Skip to main content

Full text of "Ptolemy's maps of northern Europe, a reconstruction of the prototypes"

See other formats


^B 735 32b 



L 1 











Copyright 1917 by Gudmund Schiitte. 



G^^ IS h/ 


IN 191 1, the Carlsberg Fund granted to the author of the present 
work a subsidy for the elaboration of a work dealing with Danish 
geography from an ethnic point of view, to be published by the Danish 
Society for Teutonic Philology (Selskab for germansk Filologi). In 19 1 2, 
however, this Society ceased to exist, the Royal Danish Geographical 
Society taking its place as editor of the work. At the same time, the 
subject was altered so as to embrace the Ptolemaic geography of northern 
Europe, while the subsidy granted for the elaboration was employed 
for the publication. 

The author's studies have also been subsidised by the Royal Danish 
Academy of Sciences, 

The blocks of most of the accompanying figures have been kindly 
lent by the Editor of the Scottish Geographical Magazine. 







§ I. A Brief Survey of the Manuscript Problem - i 

§ 2. Ptolemy's Predecessors in the First Century A. D - 12 

§ 3. Marinus, Ptolemy's Immediate Predecessor - 13 

§ 4. Ptolemy's Lifetime, Importance, and Principles - 14 

§ 5. Ptolemy's Successors » - 18 

§ 6. Misreadings of Latin Words - 20 

§ 7. Misreadings of Barbarian Names - 23 

§ 8. The "Milieu" as Key to Interpreting Distorted Barbarian Forms - 25 

§ 9. The Case of Metathesis - 27 

§ 10. The Case of Apocope - 28 

§ II. The Case of Parasitical Additions - 29 

§ 1 2. The Case of Onomatic Disguise - 30 

§13. The Case of Making Fictitious Repetitions - 32 

§ 1 4. The Case of False Identification - 37 

§ 15. Theoretical Arrangements - 40 

§ 16. The Question of Prototypes - 42 

§ 17. Synopsis of Prototypes - 45 

§ 18. Collective Prototype A = Europe and Environs - 48 

§ 19. Local Prototype A =. Germania - 67 

§ 20. Local Prototype Aa = North-western Germania, Chersonesus Cimbrica, and 

Scandia - 72 

§ 21. Local Prototype Ab = South-western Germania - 83 

§ 22. Local Prototypes Ac, Ad &. Ae =: Dacia and Environs - 84 

§ 23. Local Prototypes Bi &. B2 =^ the Mercantile Road from the Danube to 

the Mouth of the Vistula - 88 

§ 24. Local Prototype C = Western Gaul, Belgium, and North-western Germania - 100 

§ 25. Local Prototype D = Swabian Tribes about the Elbe - 107 

§ 26 Collective Prototypes £ & F = Eastern Germania, Sarmatia Europaea & 

Asiatica, and Scythia 112 

§ 27. Local Prototype Sk = the Scandinavian Peninsula - 127 

§ 28, The Position of the Cimbric Chersonese and the Scandian Islands within 

the Ptolemaic Germania - 1 38 

§ 29. Conclusion - 139 


A. § 30. Additions to ^ 19, Prototype J - 141 

B. § 31. Additions to § 22, Prototypes Ac, Ad, Ae - 142 



§ 32, Introduction p. 144 

§ 33. Editions of Ptolemy's Geography - 144 

§ 34. Editions of the Ptolemaic Atlas, and of Single Ptolemaic Maps - 145 

§ 35. Researches Dealing with Ptolemy or Based upon his Statements - 146 

a. Researches Dealing with Ptolemy in a More or Less General Sense ... - 146 

b. Geographic or Ethnographic Compendia, etc • 149 

c. Topography of the Cimbric Chersonese - - 1 50 


Fig. I. Ptolemaic Prototypes in Northern and Middle Europe, General Synopsis. 

2. Germania, the Cimbric Chersonese, and Scandia. Version A (Cod. Urbinas 82). 

3. — - — — - — — B ( - Burney iii). 

4. — - — — - — Type of the Roman editions, 

designed by Donis. 

5. Prototype A, Germania, according to L. Schmidt. 

6. Cimbric Chersonese and Scandia, according to the Cod. Urbinas 82, 

7. Prototype Aa-^ North-western Germania, the Cimbric Chersonese, and Scandia. 

8. Prototype Ad; South-western Germania. 

9. The Limes Transrhenanus. 

10. South-western Germania according to the Tabula Peutingeriana. 

11. Comparison of details surrounding the Vallum Hadriani. 

12. Dacia according to the Cod. Urbinas 82. 

13. Prototype Ac^ Dacia, compared with a modern map. 

14. Prototypes Ad and Ae-^ Dacia. 

15. The Ptolemaic Dacia, compared with Fig. 16, 

16. Dacia, according to the Tabula Peutingeriana. 

17. The Ptolemaic names of Dacian tribes and places redistributed according to their 
presumed correct localisations. 

18. Surviving ancient names in Dacia. 

19. Prototypes Bi and B^; the mercantile road from the Danube to the mouth of 
the Vistula. 

20. Prototypes Bi and B^^ compared with a modern map. 

21. Prototype C; Western Gaul, Belgium, and North-western Germania. 

22. Belgium and North-western Germania according to the Tabula Peutingeriana. 

23. Ancient Belgium and North-western Germania with the names from Prototype C. 

24. Prototypes £ and £; Eastern Germania, Sarmatia Europaea & Asiatica, and Scythia; 
comparison of duplicates. 

25. Prototype £ from the Cod. Urbinas 82. 

26. North-eastern Germania and Western wSarmatia with the names of the Prototypes 
£ and F. 

27. Prototype Sk; Scandia. 

28. The demarcation of Germania according to some modern representations. 

29. The Cimbric Chersonese and Germania according to the Cod. Athous Vatopediensis. 

30. A rectified Ptolemaic map of nationalities. 

31. A reconstructed map of nationalities in Ancient Middle Europe. 


PTOLEMY'S Geography, and the "Germania" of Tacitus, form the 
main foundation ^ of our knowledge concerning the barbarian north 
of Europe in classical times. It might be taken for granted that such 
extremely important documents and their sources had long ago been 
seriously examined. But the Ptolemaic description of northern Europe 
is still practically a "terra incognita". 

The present book is an attempt to supply the wanted research. It 
is based upon studies which have been carried on for many years. 

Our principal investigations concerning the different prototypes of 
Ptolemy's maps were already made 20 years ago, so the publication can 
scarcely be called precipitate. 

Of course, we do not pretend to have solved one half of the riddles 
offered by our complicated problem. If Ptolemy's Geography were to 
be examined thoroughly, it would take a lifetime, but as we have made 
some observations which at any rate shed a new light on several points, 
we thought it wiser to make an end of hesitation. For even if further 
delay might have led to still better results in certain details, the study 
will be more profitably advanced by subjecting our preliminatory obser- 
vations to revision by expert critics. 

The publication of a provisional study may possibly still be objected 
to by scrupulous philologists, but the undertaking certainly assumes a 
very different appearance, when we regard it from the geographical or 
ethnological point of view. 

Geographers and ethnologists, far from fearing the absorbing philolo- 
gical problem, have used Ptolemy's work as the foundation for large 
reconstructions, and still do so. We may name numerous publications 

XII Ptolemy's maps of northern Europe 

from later years, containing either entire reconstructions of Ptolemy's 
Atlas, or detailed statements based upon his work. E. g.: 

Miillenhoff's "Deutsche Altertumskunde", vol. II, with map designed 

by H. Kiepert 1887, republished 1906. 
Gerland, "Atlas der Volkerkunde" (Berghaus, "Physikal. Atlas", 3. ed.) 

Perthes, "Atlas antiquus", by A. v. Kampen 1892, 9*^ ed. 1916. 
Miiller's edition of Ptolemy, vol. Ill, atlas, 1901. 
V. Erckert, "Wanderungen und Siedelungen", 1901. 
Meyer's "Konversationslexikon", map of Germania designed by K. Wolff, 

M. Schonfeld, "Worterbuch der altgermanischen Personen- und Volker- 

namen", in Streitberg's "Germanische Bibliothek", 191 1. 
R. Kiepert, "Formae orbis antiqui"; e. g. reconstructed Ptolemaic map 

of Europe (191 1) and map of Germania (1914). 

We may specially mention the latest publications of ethnological 

Caspar Zeuss, "Die Deutschen und die Nachbarstamme", ist ed. 1837, 
republished 1903 (unaltered). 

O. Bremer, "Ethnographie der germanischen Stamme" in Paul's monu- 
mental manual "Grundriss der germanischen Philologie", 1899, 
republished separately 1905. 

In all these publications, Ptolemaic data are used as a basis without 
any serious attempt to solve the philological problem. In order to prevent 
scientists from continuing such a proceding, it is not merely allowable, 
but necessary to publish the results of a research in which the attempt 
is at any rate made, — whether the outcome is satisfactory or not. 

The necessity of revising the traditional ideas about classical geography 
is specially urgent within the region of the author's native country, i. e. 

Although Ptolemy's work offers an attractive base for such a study, 
it has, since the middle of the 19th century, been lamentably neglected. 
This neglect principally concerns the much discussed problems, as to 


whether the classical Cimbri, Charudes, and AngHi, are to be identified 
with the modern Jutlandic populations of Himmerboer, Hardboer, and 
Angelboer, — or whether they are to be placed somewhere in Germany 
south of the Elbe. Of late years, several authors have published very 
learned researches dealing with the matter, e. g. in Germany Mullenhoff, 
in Sweden Erdmann, in England H. M. Chadwick and R. W. Chambers. 
But none of these authors has ventured upon examining the prototypes 
of Ptolemy's map in detail. As such important problems concerning 
the past of the Danish and English peoples could not be treated in a 
satisfactory manner, while Ptolemy's map remained an unexplored laby- 
rinth, we subjected classical Jutland to a special study, and this became 
the nucleus of the present work. 

Originally, it was our aim to write a compendious introduction con- 
cerning the question of text criticism. We also published some provi- 
sional sketches in "The Scottish Geographical Magazine", February and 
June 1 91 4, and in Paul & Braune's "Beitrage zur Geschichte der deut- 
schen Sprache und Litteratur", Vol. 41, 191 6, at the same time anti- 
cipating our main results concerning the prototypes of Ptolemy's Atlas. 
Cf. our article in the "Saga Book of the Viking Soctety", 191 3, Vol. VIII, 
part I, and in the "Mitteilungen zur Geschichte der Medicin und der 
Naturwissenschaften", 1914, Vol. XIII, No. 5. 

On further consideration we found it inadviseable to publish in one 
volume a detailed MS. criticism and a detailed investigation of carto- 
graphic prototypes. The problem of text criticism is so complicated 
as to require a separate volume. After being introduced into this dange- 
rous labyrinth, the reader would scarcely retain sufficient energy to 
venture upon the equally absorbing task of tracing Ptolemy's cartogra- 
phic scheme. 

In addition, the state of general European warfare prevented us from 
carrying on our text studies in the countries where the Ptolemaic MSS. 
are preserved. 

We therefore resolved to publish our studies of Ptolemaic text criti- 
cism occasionally, whereas we limit the present volume to the carto- 
graphic problem. It will merely be introduced by a paragraph which 
briefly sums up the main points of the text question. 

Fortunately enough, a lately discovered MS., the Urbinas 82, pre- 

XIV Ptolemy's maps of northern Europe 

serves the Ptolemaic atlas in a state which must be called excellent. 
Trusting the evidence of the greatest Ptolemaic MS. experts, — e. g. 
Prof. Jos. Fischer — we have based our studies firstly and mainly on 
this document which outweighs most other existing representatives of 
the famous classical geographer's work. Critics may object to our pro- 
ceding^ but it is at any rate a simple and practically justifiable expedient 
during the present difficult conditions of text research. 

Readers of our previous articles will notice that our theories have 
in some respects undergone a radical revolution. The complete reversal 
of some theses may seem startling and at the first sight cause the 
impression of ''vestigia terrent". — In an unexplored field of study it is, 
however, impossible for a pioneer to avoid some serious mistakes. Any 
conceivable possibilities must be taken into account, simply for argu- 
ment's sake. A number of them which have at first seemed satisfactory 
will, in the long run, prove misleading, but yet they have fulfilled a 
mission, namely that of contributing to the exhaustive discussion of our 

The term "Gothonic" is in this work used instead of the synonymes 
"Teutonic" and "Germanic" which are unpractical because of their ambi- 
guity. Cf. Th. de la Saussaye, "The Religion of the Teutons" p. 79, 
— and Axel Olrik, "Arisk og Gotisk" ("Danske Studier" 1916). 
"Germanic" which is nowadays adopted by several English scientists, is 
especially bad, for it has no less than 11 or 12 difierent significations, and 
the English substantive "Germans" can only mean "inhabitants of Ger- 
many"^). Cf. our treatise "Gothonic Names" in the "Publications of the 

') We have only noticed two exceptions, namely Chambers, "Widsith", where the 
Scandinavians are called "North-Germans", etc. (p. 157); and Stjerna's "Essays on Heowulf", 
transl. by Clark Hall. 


Society for the Advancement of Scandinavian Study", December 191 2 
(Urbana, Illinois), and our article "Germaner" in R0rdam's "Illustreret 
Konversations-Lexikon" (Hagerup), where the different significations are 
pointed out. — The term "ethel Gotena" — "nobility of Goths" — is 
used already in Old English in order to express the flower of the Teu- 
tonic heroes, see Widsith, part III. In the Edda, and in other Old Norse 
traditions, "Gotnesk" was equivalent to "Gothonic", "Teutonic"; and 
"Got-thiod", i. e. "Gothic Nation", meant the whole of the Teutonic 
group. Cf. W. Grimm, "Deutsche Heldensage", 3rd. ed. p. 6: "Sehr 
natiirlich hat die Edda hernach gothisch im allgemeineren Sinne genom- 
men". — "Gothic" was used in the same collective sense by Icelandic, 
English, Dutch, Swedish, Norwegian, and Danish scientists from the 
1 6th to the 19th century. This nomenclature is mentioned e. g. in the 
"Encyclopedia Brittannica", 9th ed., 1876^). — We prefer the longer 
form "Gothonic" in order to avoid ambiguity. 

We owe much valuable information to Professor J. L. Heiberg in 
Copenhagen, the editor of Ptolemy's "Syntaxis". 

^) Art. "English Language", by J. A. H. Murray, p. 391. "The Angles, Saxons and 
their allies belonged to the Teutonic or Gothic branch of the Aryan family, represented 
in modern times not only by the English and their colonies, but by the populations of 
Germany, Holland, Denmark and the Scandinavian peninsula .... For more than looo 
years, the Teutonic or Gothic stock has been divided into three branches." 

Art. "Goths", by E. A. Freeman, p. 847. "The name came . . to be used as a 
philological or ethnological term; we heard of "Gothic nations", "Gothic languages" etc., 
meaning "Teutonic" in the widest sense. The name was also first scornfully, then respect- 
fully, applied to a style of architecture which has some claim to be called Teutonic as 
opposed to Greek or Roman, but which has nothing whatever to do with the Goths as a 

The name "Gothic", meaning "Teutonic", is also mentioned in the i ith edition, art. 
"Teutonic Peoples" by H. M. Chad wick, p. 679. 

XVI Ptolemy's maps of northern Europe 

We are especially indebted to the authority in the study of Ptolemy's 
MS. atlas, Professor Jos. Fischer S. J. in Feldkirch, who has most 
liberally allowed us to use his large material of MS. reproductions and 
whose information and suggestions have been a great help. We there- 
fore dedicated the present volume to him, hoping that our theories may 
in some points contribute to the advancement of the highly interesting 
study which has been so greatly promoted by his efforts and achieve- 


The study has until now been handicapped by the fact that the 
critics would not acknowledge the atlases of the Ptolemaic MSS. as 
directly derived from the author's original cartographic work. These 
atlases were regarded as reconstructions from the MS. text, executed 
possibly by the Alexandrine grammarian Agathodaemon in the 5th 
century, or even later, and consequently deemed unworthy of consi- 

We may name some of the critics who more or less distinctly share 
this view of the MS. atlas. 

Fabricius, "Bibliotheca Graeca", III, p. 414. 

Heinrich Kiepert^ "Lehrbuch der alten Geographic", 1878, pag. 10.^) 

Berger, "Geschichte der griechischen Erdkunde". 

— , "Die Grundlagen des Marinus-Ptolemaischen Erdbildes" (Berichte 
d. phil. hist. CI. d. sachsischen Gesellsch. d. Wissenschaften". 1898, 

p. 87-143)'). 

Christ, "Geschichte der griechischen Literatur" (in Miiller's "Handbuch 

der class. Altertumskunde", VII, p. 506), 1888. 
Henry Zondervan, "Allgemeine Kartenkunde", 1901. 

It may be added that the Russian scholar Kunik wrote to Kiepert 
on Jan. 7th 1892, directly drawing his attention to the atlas in the 
phototypic reproduction of the Athos MS., published by Sewastionow 
and Langlois in 1867. Kunik had noticed the great difference between 
this atlas and the reconstructed maps in Kiepert's Atlas antiquus and 

^) In order to avoid misunderstanding^ we may quote what K. Kretschmer says about 
Kiepert's and Berger's opinions, "Zeitschrift der Gesellschaft fiir Erdkunde zu Berlin", 
1913, Heft 10, S. 28, He states, "dass H. Kiepert und H. Berger keineswegs den vor- 
herigen Entwurf von Karten von seiten des Ptolemaus in Abrede gestellt haben; im Gegen- 
teil, H. Kiepert sagt (Lehrb. S. lo) ausdriicklich, dass die Konstruktion der Karten dem 
daraus erst abstrahierten Text vorangehen musste. H. Berger bestreitet vielmehr, dass 
Ptolemaus die zuvor konstruierten Karten seiner "Geographie" als Illustration beigegeben 
habe, Ptolemaus wollte absichtlich keine Karten liefern". 


thence concluded iJiatKiepert' had either been unaware of the reproduc- 
tion or that he — as an accurate critic — had put it aside on purpose, 
deeming it to be of Uttle practical value ^). Kunik now wanted to know 
whether the Athos Atlas might be regarded as truly Ptolemaic or not. 

Kiepert's answer is unknown, Roediger adds, but the later editions 
of his atlas do not seem to betray that he has in any way altered his 
previous opinions concerning the MS. atlases of the Ptolemaic Geography. 
Nor is any trace of an altered scheme to be found on the map of 
ancient Europe, designed by his son and editorial heir R. Kiepert in 
191 1 ("Formae orbis antiqui"). 

Thus the systematic ignoring of the Ptolemaic MS. atlases is shared 
by almost all scholars, including the latest editors of the text such as 
Wilberg 1838, Miillenhoff 1873, and C. Miiller 1883— 1901. 

As late as 191 4, K. Kretschmer finished an article thus: "We con- 
clude that the MS. maps do not originate directly from Ptolemy, but 
at the best from Agathodaemon who lived after him"^). 

A. Herrmann later has taken up Kretschmer's point and finally 
maintains: "One result has proved certain, — our basis is not formed 
by the MS. maps but by the eight books of the text. Only the 
exterior qualities can be illustrated by means of the atlases : they supply 
information concerning the number of maps designed by Marinus, and 
concerning the regions described by him, and they show the technical 
means by which the graduation and the mountains, rivers, and towns 
were represented. But wherever the positions of the points described 
and the forms of names are concerned — and that is finally our prin- 
cipal subject — the text and not the maps must be our guide". ^) 

We may point out some principal arguments of Kretschmer and 
other critics who maintain that Ptolemy is not the author of the MS. 

*) The letter is reprinted by Roediger in the Preface to the second volume of Miillen- 
hoffs "Deutsche Altertumskunde", p. XV. Cf. the following sentences (our italics): 

"Zu meiner Ansicht iiber die Welten war ich nach wiederholter Prufung des Textes von 
Ptolemaus gelangt", Kunik writes. "'Erst vor einigen Tagen kam es mir in den Sinn, die 
Karten zu befragen, welche im Athosmanuscript des Ptolemaus enthalten sind und von 
Sewastionow photographiert wurden (Geographic de Ptolemee, Reproduction photo-litho- 
graphique; Paris, Didot 1867). Ich wurde stutzig, als ich Karte LXXVI mit der Ihrigen 
verglich und kam endlich dazu, vorauszusetzen, dass Sie entweder die wenig verbreitete, 
teure Ausgabe von 1867 nie zu Gesicht bekommen, oder dass Sie als feiner Kri- 
tiker die Karten als wenig brauchbar bei Seite gelassen haben . . . Bei dieser 
Lage der Dinge halte ich es fiir das Beste, meine Zuflucht zu Ihnen zu nehmen, indem 
ich Sie um gtitige Aufklaring iiber die Athoskarte No. 76 bitte. Darf man sie als eine 
Copie der von Ptolemaus selbst entworfenen Karte ansehen?" 

*) "Die Ptolemauskarten", in "Petermanns Mitteilungen", 19 14, p. 142., cf. Kretschmer's 
statements in the "Zeitschrift des Vereins fiir Erdkunde zu Berlin", 1913, Heft 10. 

^) "Marinus, Ptolemaus und ihre Karten", in "Zeitschrift der Gesellschaft fiir Erdkunde zu 
Berlin", 1914, No. 10. 


atlases, cf. Dinse in "Zeitschr. d. Vereins f. Erdkunde zu Berlin", 19 13, 
p. 745 seq. 

Firstly, the conclusion is drawn from Ptolemy's own words in his 
Geography I, XVIII, 2, stating that repeated copying would always tend 
towards the deterioration of the maps. It is supposed that Ptolemy 
would in order to prevent such deterioration publish his geography in 
tabular form without maps. 

Secondly it is urged that Ptolemy has in his geography laid stress on 
the conic projection as preferable to the cylindric, — but the atlases 
contain only one specimen of the former; the remainder are square maps, 
designed in the cylindric projection which was by Ptolemy characterized 
as inferior. 

Thirdly, none of the MS. atlases are by the copyists attributed 
directly to Ptolemy, nor are maps designed by Ptolemy mentioned any- 
where in classical or mediaeval literature. On the contrary, several MS. 
atlases contain a notice attributing them to "Agathos Daimon", a mechanic 
in Alexandria. This author has again been identified with an Alexan- 
drine grammarian Agathodaemon who lived in the 5th century A. D. 
— The authorship of Agathodaemon has been regarded as most con- 
clusive, and declared to be quite irreconciliable with the assumption that 
the Ptolemaic MS. atlases could have been designed by Ptolemy himself. 

The Ptolemaic MS. atlases have already been defended against the 
sceptics in 1822 by N. H. Brehmer, "Entdeckungen im Altertum", 
Heft I, p. II, and in 1828 by Heeren, "De fontibus geographicis 
Ptolemaei" ("Comment. Gotting." VI, p. 66). 

But it was not before the beginning of the 20th century that a more 
general reaction against the scepticism made itself felt. 

Prof. Jos. Fischer S. J. in Feldkirch is the main upholder of the 
revised theory recognizing the better MS. atlases as true continuations 
of Ptolemy's own work. Whereas his predecessor C. Miiller has made 
the greatest collective study of the context, Fischer has undertaken a 
corresponding collection of the MS, atlases in photographic reproduction, 
originating from more than 40 Ptolemaic MSS. The collection has been 
supported by the ''Istituto Austriaco di studii storici". Fischer's provi- 
sional results are principally found in the treatises "Die handschriftliche 
Ueberlieferung der Ptolemaus-Karten" 19 12, and "An Important Ptolemy 
Manuscript" 191 3. 

A report of Fischer's as yet unpublished results together with 
numerous inividual observations is given by Paul Dinse, "Die hand- 
schriftlichen Ptolemauskarten" ("Zentralblatt f. Bibliothekswesen" XXX, p. 


results by a series of comparisons with the Tabula Peutingeriana and the 
insignia in the Notitia Dignitatum. 

A regular scale of development may be observed, stage I with few 
pictorial elements and no living beings, stage II with a growing number 
of pictorial elements among which are some few living beings in repose, 
stage III with complete overgrowth of pictorial elements among which 
several living beings in movement. Within this perspective, the Ptolemaic 
MS. atlases distinctly occupy the oldest stage, whereas all other existing 
documents, dating from the 4th, 5th, and 6th cent. A. D., represent 
later developments. Cf. our treatises in "The Scott. Geogr. Mag.", Febr. 
and June 1914, and in the "Mitteil. z. Gesch. d. Medicin u. d. Natur- 
wiss.", 1914, Vol. XIII, No. 5. 

Our third question concerns the additional details — lines, vignettes, 
spellings and entire names — which do not occur in the Ptolemaic text. 
The figures of longitude and latitude leave sufficient room for individual 
variation, — e. g. Ptolemy represents rivers and mountain chains gener- 
ally by the two terminal points only, whereas the lines between these 
extremities are left to the cartographer's divination. A reconstructor 
with a lively fancy might here introduce quantities of "naturalistic details" 
without directly destroying the traditional framework, as sometimes occurs 
in late mediaeval MSS. The older MS. atlases, however, do not betray 
the slightest inclination of the cartographers towards using their liberty 
in an arbitrary manner. They represent features, it is true, which are 
not implied by the words of the text, but such additions are made on a 
limited scale and characterized by no divergence from the general scheme 
of the Ptolemaic work. We notice e. g. that the mountains and rivers 
of western Germany, as given by the Cod. Urbinas 82, are derived from 
a special map of Roman fortification lines, cf. § 21. — Another addition 
to the Ptolemaic scheme is the more specialised classification of towns: 
whereas Ptolemy distinguishes only two classes, the atlases add a third, 
as stated directly in the Editio Romana 1478: "Urbes insignes, secunde 
urbes, tercie urbes" ^). The possibility is perhaps not excluded that the 
more detailed classification may have been a mediaeval addition, but 
there are no obvious reasons supporting this suggestion and the distri- 
bution of classes II and III seems to point strongly towards tradition 
from ancient times. — Finally, we notice that the MS. atlases contain 
sometimes the more correct spelling or give entire names. which are left 
out in the text. 

Our main result may be expressed by the words of J. Fischer cited 
above with special reference to the maps of the Cod. Urbinas 82: ''they 
.... represent the maps designed by Marinus". 

^) Cf. J. Fischer, "An Important Ptolemy MS.*', in the "Catholic Hist. Records and 
Studies", New York, 1913, p. 227. 


As a matter of fact, the possibility — or even likelihood — of this 
explanation is admitted by those scholars who have lately denied Pto- 
lemy's authorship. Kretschmer says in "Petermanns Mitteilungen", 191 4, 
p. 142: "We cannot sans phrase deny the possibility that the maps in 
their fundamental elements may be traced back to ancient times and that 
they, like the text, have been preserved by steadily repeated copying". 
"Nobody denies . . . that Ptolemy must have constructed a cartographic 
prototype on the basis of the map of Marinus". Herrmann says, "Zeit- 
schr. des Vereins f. Erdkunde zu Berlin", 1914 (Heft 10): "If we con- 
sider how much Ptolemy — even when attacking Marinus — depends 
upon the latter, we must take it for granted that those 68 maps for 
which the text gives instructions as to the method of design, are in 
reality nothing else but the maps of Marinus." 

If this is admitted by the opponents, the reasons for further ignoring 
the Ptolemaic MS. atlases have practically been abandoned. 

It remains but to add some few words concerning the MS. atlases 
regarded from the point of view of text editors. — This is one of those 
regions where the method of Ptolemy's sceptical critics appears in its 
most astonishing light. 

These expert philologists profess to give the sum total of the diver- 
gent readings, known to them. Anxious to be exhaustive, they quote 
not merely the MSS. containing the original Greek text, but also Latin 
translations, and even printed editions from the 15th and i6th centuries. 
But the readings of the MS. atlases are consistently ignored. 

In order to understand this system, we might naturally expect a chapter 
or paragraph tending to prove that the MS. atlases are later than the 
15th and 1 6th century and contain a much inferior reading than do the 
first printed editions. But no such chapter or paragraph is found. The 
readings of the MS. atlases are simply ignored sans phrase ! ! 

As the editors give no reasons, we must apply to expert palaeographers 
such as Messrs. Krumbacher, Mercati, and P'ranchi (cf. Fischer, "Die 
handschriftliche Ueberlieferung der Ptolemaus-Karten", p. 228, and a 
letter from the late Dr. BJ0rnbo, preserved in the Copenhagen Uni- 
versity Library). To our surprise we learn here that there is no diffe- 
rence of age between the MS. texts and the accompanying atlases. 
The Laufentian. XXVIII, 41, the Mediolan. Ambrosian. 527, the Urbinas 
83 and 82, the Fabritius fragm. in the Copenhagen Univ. Libr. — both 
texts and atlases — , would all have been executed about 1200, whereas 
the Athos MS. reproduced by Sewastionow and Langlois would be some 
50 years later. The first named 5 MSS. are of a distinctly superior 

Our review of the present editorial standard consequently results in 
the following somewhat startling conclusion: superior MS. readings from 

8 Ptolemy's maps of northern Europe 

the 13th century have been ignored in favour of more or less corrupt 
readings from the 15th and i6th century printed editions!! 

The discovery of such procedure cannot but gravely shake our con- 
fidence in the authority of the "expert" editors. The whole collection and 
verification of text material must be deemed not merely unsatisfactory, 
but utterly superficial. 

The bad consequences of such false methods can soon be pointed 
out in detail. 

The editions leave out names which are found in the atlases. 
According to Fischer, we miss e. g. Karkum, which is in the Urbinas 
82 mentioned as an additional name of the town Babilon i Egypt. The 
possibility that the addition could be of mediaeval origin is excluded be- 
cause the vernacular Egyptian name Karkum vanished at the close of the 
Roman period. 

Moreover, Ptolemy is repeatedly accused of corruptions which could 
have been amended by the aid of the atlases. 

E. g. the Ptolemaic name of the present Tongern is given as Atu- 
akuton, and the corrected form Atuatukon is added "e conjectura". But 
the atlases of the Codd. Mediolan. Ambrosian. 527 and Urbinas 83 
quite clearly write Atuatokon, which is consequently the true Ptolemaic 

In eastern Germania, the editions record a town Setuia. But the one 
class of MS. atlases write the name Artekuia (or Artekvia), and we shall 
show later on that an addition of both readings gives the correct 
Ptolemaic form *Arsekuia which is in reality a duplicate of the neigh- 
bouring Arsikua (or Arsikva). The evidence of the Artekuia-class of 
MSS. is highly valuable, as it unveils a sample of Ptolemy's well known 
duplicates, pointing towards the lost prototypes of his work. Without 
the help of the MS. atlases we should never have recognized Setuia as 
duplicate of Arsikua. 

The above consideration radically alters the valuation of the material 
for examining the Ptolemaic cartography. 

This altered view would still be of relatively little import, if our aim 
were to analyze Ptolemy's work in its most minute local details. Then 
we should still be obliged to fix the position of any mountain, river 
or town by means of the longitude and latitude indicated in the text, 
and we should have to discuss the complicated questions of text 
genealogy in order to make our choice between the divergent figures. 

We do not however aim at such gigantic research. The results would 


hardly be worth the trouble, — at least so far as Germania or Sarmatia 
are concerned — , for in these and other parts of the ancient barbarian 
world, one half of Ptolemy's "exact" astronomic definitions are pure 

Our task is only to furnish some preliminatory observations, in order 
to prepare a methodical investigation of Ptolemy's lost prototypes from 
a cartographic point of view. And for this purpose, the hitherto acces- 
sible material seems to be sufficient. 

We agree with C. Miiller and Mommsen who state that the Codex 
Vaticanus 191 is the most valuable of all context MSS. Cf. the treatises 
of the two said authors in the periodical "Hermes", Vol. XV. 

As tp the MS. atlas, the approximate agreement of its best repre- 
sentatives may be regarded as a trustworthy guide. 

According to Jos. Fischer, the MS. atlases are divided into two ver- 
sions, one with 27 maps, and one with some 6S. The 27 version corre- 
sponds to C. Miiller's "Byzantine Family" of context MSS., and its 
main representatives are the Cod. Urbinas 82, the Cod. Athous Vato- 
pediensis (Athos Atlas), and the Venetus Marcianus 5661). The 68 version 
corresponds to Miiller's "Asiatic P'amily", and its main representatives 
are the Laurentianus Pluteus XXVIII, 79, the Mediolanus Ambrosianus 
527, the Urbinas 83, and the Burney 111, 28. 

The Athos Atlas has been published in phototypic reproduction by 
Sewastionow and Langlois, Paris 1867. The Urbinas 82 will soon be 
reproduced by Jos. Fischer. 

The fact that only the Athos copy of the MS. atlas has hitherto 
been reproduced, caused us first to use this document as a cartographic 
basis of our investigations. We attached considerable value to the fol- 
lowing details: i. the design of German mountains; 2. the representation 
of the river Loire (as touching the city of Orleans); 3. the representation 
of Scandinavian coasts; 4. the colouring, separating the Cimbric Cher- 
sonese and the Scandian islands from Germany; 5. the occurrence of a 
duplicate of the name Asanka in Bohemia. Cf. our paper "Une carte 
du Danemark, agee de 1900 ans", in the periodical "Le Danemark" 
Nov. 1912. 

Later, we were informed by Jos. Fischer that the reproduction of the 
Athos Atlas is all but reliable, and that the original MS. itself is exe- 
cuted in a careless manner, forming no solid basis for conclusions. As 
to the duplicate of Asanka, it is not found in any of the other MS. 
atlases and consequently cannot be regarded as truly Ptolemaic. 

Thus we had to discard a series of wrong presumptions and to accept 
rather the Codex Urbinas 82 as our principal basis. 

^) Cf. C. Mailer's treatise in the "Archives des missions scientifiques et litteraires", 1867. 

8 Ptolemy's maps of northern Europe 

the 13th century have been ignored in favour of more or less corrupt 
readings from the 15th and i6th century printed editions!! 

The discovery of such procedure cannot but gravely shake our con- 
fidence in the authority of the ''expert" editors. The whole collection and 
verification of text material must be deemed not merely unsatisfactory, 
but utterly superficial. 

The bad consequences of such false methods can soon be pointed 
out in detail. 

The editions leave out names which are found in the atlases. 
According to Fischer, we miss e. g. Karkum, which is in the Urbinas 
82 mentioned as an additional name of the town Babilon i Egypt. The 
possibility that the addition could be of mediaeval origin is excluded be- 
cause the vernacular Egyptian name Karkum vanished at the close of the 
Roman period. 

Moreover, Ptolemy is repeatedly accused of corruptions which could 
have been amended by the aid of the atlases. 

E. g. the Ptolemaic name of the present Tongern is given as Atu- 
akuton, and the corrected form Atuatukon is added "e conjectura". But 
the atlases of the Codd. Mediolan. Ambrosian. 527 and Urbinas 83 
quite clearly write Atuatokon, which is consequently the true Ptolemaic 

In eastern Germania, the editions record a town Setuia. But the one 
class of MS. atlases write the name Artekuia (or Artekvia), and we shall 
show later on that an addition of both readings gives the correct 
Ptolemaic form *Arsekuia which is in reality a duplicate of the neigh- 
bouring Arsikua (or Arsikva). The evidence of the Artekuia-class of 
MSS. is highly valuable, as it unveils a sample of Ptolemy's well known 
duplicates, pointing towards the lost prototypes of his work. Without 
the help of the MS. atlases we should never have recognized Setuia as 
duplicate of Arsikua. 

The above consideration radically alters the valuation of the material 
for examining the Ptolemaic cartography. 

This altered view would still be of relatively little import, if our aim 
were to analyze Ptolemy's work in its most minute local details. Then 
we should still be obliged to fix the position of any mountain, river 
or town by means of the longitude and latitude indicated in the text, 
and we should have to discuss the complicated questions of text 
genealogy in order to make our choice between the divergent figures. 

We do not however aim at such gigantic research. The results would 


hardly be worth the trouble, — at least so far as Germania or Sarmatia 
are concerned — , for in these and other parts of the ancient barbarian 
world, one half of Ptolemy's "exact" astronomic definitions are pure 

Our task is only to furnish some preliminatory observations, in order 
to prepare a methodical investigation of Ptolemy's lost prototypes from 
a cartographic point of view. And for this purpose, the hitherto acces- 
sible material seems to be sufficient. 

We agree with C. Miiller and Mommsen who state that the Codex 
Vaticanus 191 is the most valuable of all context MSS. Cf. the treatises 
of the two said authors in the periodical "Hermes", Vol. XV. 

As tp the MS. atlas, the approximate agreement of its best repre- 
sentatives may be regarded as a trustworthy guide. 

According to Jos. Fischer, the MS. atlases are divided into two ver- 
sions, one with 27 maps, and one with some 68. The 27 version corre- 
sponds to C. Miiller's "Byzantine Family" of context MSS., and its 
main representatives are the Cod. Urbinas 82, the Cod. Athous Vato- 
pediensis (Athos Atlas), and the Venetus Marcianus 5661). The 68 version 
corresponds to Miiller's "Asiatic Family", and its main representatives 
are the Laurentianus Pluteus XXVIII, 79, the Mediolanus Ambrosianus 
527, the Urbinas 83, and the Burney 111, 28. 

The Athos Atlas has been published in phototypic reproduction by 
Sewastionow and Langlois, Paris 1867. The Urbinas 82 will soon be 
reproduced by Jos. Fischer. 

The fact that only the Athos copy of the MS. atlas has hitherto 
been reproduced, caused us first to use this document as a cartographic 
basis of our investigations. We attached considerable value to the fol- 
lowing details: i. the design of German mountains; 2. the representation 
of the river Loire (as touching the city of Orleans); 3. the representation 
of Scandinavian coasts; 4. the colouring, separating the Cimbric Cher- 
sonese and the Scandian islands from Germany; 5. the occurrence of a 
duplicate of the name Asanka in Bohemia. Cf. our paper "Une carte 
du Danemark, agee de 1900 ans", in the periodical "Le Danemark" 
Nov. 1912. 

Later, we were informed by Jos. Fischer that the reproduction of the 
Athos Atlas is all but reliable, and that the original MS. itself is exe- 
cuted in a careless manner, forming no solid basis for conclusions. As 
to the duplicate of Asanka, it is not found in any of the other MS. 
atlases and consequently cannot be regarded as truly Ptolemaic. 

Thus we had to discard a series of wrong presumptions and to accept 
rather the Codex Urbinas 82 as our principal basis. 

^) Cf. C, Miiller's treatise in the "Archives des missions scientifiques et litteraires", 1867. 


But this changed valuation of MSS. has not altered our theories con- 
cerning the assumed Ptolemaic prototypes. As a matter of fact, the 
change was but little, because the Athos Atlas and the Urbinas 82 belong 
to the same group of MS. atlases, the version with the 27 maps. 

Generally speaking, our reconstructions of prototypes remain unaffected. 
The doubts concerning the reading of several names are scarcely of any 
import to these theories. 

Far from fearing that new discoveries within the text study will shake 
our prototype theories, we believe rather that the latter will prove a 
practical means of ascertaining the preferable texts. 


The political centralisation of the classical world within the Roman 
Empire led directly to a corresponding centraUsation of the geographical 
and statistical studies. About the beginning of the Christian era, great 
activity was displayed in chronicling the sum total of acquired know- 
ledge, both from the well known Mediterranean shores, and from the 
recently conquered reigns in the far North and East. 

The Imperial family played an important part in this activity. 

M. Vipsanius Agrippa, the son-in-law of Augustus, wrote statistical 
"Commentaries" and designed a map of the world which was finished 
between 27 and 20 B. C. 

The Emperor Augustus himself also contributed greatly to the or- 
ganisation of statistical and geographical studies. It is well known from 
the Bible that he arranged the first world-census in Europe ; this occurred 
in the birth-year of Christ. Seven years previously, a revision of Agrippa s 
map of the world had been undertaken in Rome by order of Augustus. 
The Imperial map thus constructed was of colossal size and painted in 
bright colours. Copies seem to have been placed in several provincial 

The classical geographers Strabo and Pliny are our main authorities 
concerning the above-mentioned undertakings of Agrippa and Augustus. 
Cf. the special literature, quoted by O. Bremer in his "Ethnographic der 
germanischen Stamme" § 6. 

The Imperial publication became the foundation of all subsequent 
maps of the world during the remaining period of antiquity and during 
the whole of mediaeval times. 


According to Ptolemy's Preface, his geography and atlas were directly 
based on a work of Marinus from Tyrus, This scholar, as an older con- 

§ 3- MARiNus, Ptolemy's immediate predecessor ii 

temporary of Ptolemy, must have lived in the first half of the second 
century A. D. Death overtook him before completing his work. 

Ptolemy in ch. VI of the Preface characterizes his predecessor's work 
with the following words: 

*'Marinus from Tyrus appears to be the last of our contemporaries who 
carried on the study with great zeal. In addition to the older commen- 
taries which had come to our notice, he has discovered several more. 
With great accuracy, he has investigated the works of nearly all prece- 
ding authors, subjecting them to reasonable emendations". 

We agree with this statement of Ptolemy's, — as a matter of fact, 
Marinus must have been gifted with colossal energy in collecting. Ptolemy 
has only augmented his collections in some few regions, mentioned in 
Preface ch. XVII, viz.: the coasts of Africa, India and East Africa, the 
extreme East Asiatic port Cattigara, China, and the mercantile road from 
the silk-producing country to Palimbothra. 

Ptolemy, however, felt obliged to criticize the scheme of Marinus in 
several respects: the emendations introduced were not sufficient and 
especially the square projection, used for constructing the maps, was not 
up to scientific requirements, cf. Preface ch. XVIII. 

Still, as a matter of fact, the Ptolemaic maps have preserved this 
projection except one and as we have no reason for doubting their per- 
tinence to Ptolemy's age we must assume that the criticized scheme of 
Marinus remained the basis of the completed atlas. 

Taking for granted that the existing Ptolemaic geography and maps 
represent the unaltered work of Marinus, we must agree with Ptolemy's 
judgment that they betray a considerable want of critical talent. 

Marinus was not gifted with great divination in interpreting the phy- 
sical outlines of the original maps from which he constructed his own 
atlas. He often mistakes sea-coasts for rivers, and rivers for mountains, 
or mountains for tribes and so on. North is changed into west, and 
west into south, etc. 

His philological capacity was still weaker. He was completely unable 
to read and interpret barbarian names from little known regions. When 
two of his prototypes had the same name spelt a little differently, he 
did not recognize the identity. Thus the same name may occur twice, 
thrice, and even four times on the maps. 

As the maps of Marinus are now only preserved through the medium 
of Ptolemy's work, it is often difficult to distinguish to which author the 
various features are attributable. In the following research, we have 
therefore introduced the expression "the Ptolemaic constructor", as em- 
bracing both. 



Claude Ptolemy in Alexandria succeded in completing the unfinished 
work of Marinus towards the end of the second century. The publication 
of his FeoyQacpiKr] vqyrjyrjoig forms the culmination of classical geography, 
and with all its faults, it may be called the most colossal exploit ever 
achieved in geographical literature. It marks a new epoch in so far^ as 
not only the description, but also the accompanying monumental atlas 
escaped destruction and has come down to posterity. And here, in con- 
trast to the previous absolute want of cartographic relics, vast material 
for study is suddenly placed within our hands. For more than 1 500 
years, it was destined to remain unrivalled both in quantity and in 
quality. Since the beginning of the humanistic era, it dominated for 
centuries all construction of scientific maps. 

The date of Ptolemy's birth and death is not recorded. He is known 
to have undertaken astronomic observations in Alexandria during the 
reign of the Emperors Hadrianus and Antoninus, more exactly between 128 
and 151 A. D.^). As Ptolemy's Dacian tribe-names Biessoi and Sabokoi 
with their surroundings re-appear only in the "Bellum Marcomannicum" 
of Julius Capitolinus, it is possible that Ptolemy lived to witness the be- 
ginning of the war against the Marcomans which was carried on from 
166 to 180. 

Ptolemy is known as the most famous astronomer of antiquity, though 
others more truly deserved the title. 

In the Preface, he spends numerous chapters on correcting wrong' 
astronomic principles and details in the collections of his predecessor 

In his own geography, Ptolemy relates the length of the midsummer 
day at numerous important points of the world. The atlas marks the 
places of observation by means of crosses, and by vignettes with towers. 
Physical outlines and even the tiniest boroughs are localised by longitude 
and latitude, so that we may reconstruct the atlas on the base of the 
text with relative exactness. In the atlas, the lines of longitude and 
latitude are designed in the most accurate manner, cf. Dinse's description. 
The towns of the most important countries are arranged by Ptolemy 
according to their pertinence to the respective tribal districts. Singularly 
enough, all islands except Great Britain escape this sort of ethnic clas- 
sification. The atlas expresses the classification by means of ethnic 
signs 2). Statistical signs — vignettes with towers or battlements or without 

^) See Heiberg's edition of Ptolemy's "Opera astronomica minora", Index p. 271, 273. 
') The signs seem to have occurred already in some original maps, cf. § 10, but their 
systematical introduction into the atlas seems to be due to Ptolemy. 

§ 4- Ptolemy's lifetime, importance, and principles 13 

either, distinguish 3 classes of towns: the "urbes insignes, secunde and 
tercie urbes"^). 

The critical principles, enunciated by Ptolemy in the Preface, are 
praiseworthy, cf. especially chapter V. 

"From the traditions of successive ages, which we have collected, it 
appears that many inhabited parts of our Continent have still not come 
to our notice, owing to the difficulty in exploring them. Whereas others 
are not duly described according to their real appearance, owing to the 
carelessness of those who received the information. Finally, several have 
now actually changed their appearance, owing to revolutions or trans- 
formations" .... 

"The later times generally supply more accurate notice concerning 
all regions which are not fully known" .... 

"Therefore it is generally necessary to pay attention to the latest 
records of our times. In our statements, we must observe what is re- 
corded nowadays, and in ancient tradition we must discriminate between 
what is trustworthy and what is not". 

Chapter IV points out that the reports of travellers are generally to 
be placed in the first rank. In Chapter XI, the incredulity of Marinus 
against traders is criticized. 

In Book II, ch. 1., Ptolemy declares that he does not take into account 
the "mixed stuff" (to tioXvxovv) which the historians relate in describing 
the peculiarities of various nations, "except when some generally recorded 
detail requires an exact and reasonable statement". 

Such were Ptolemy's principles. If those principles were carried out 
only halfway to their aim, a splendid work must have resulted. The 
question is now, how far Ptolemy succeded. 

There can be no doubt that the mere accomplishment of a work like 
Ptolemy's was a unique achievement. 

And on several points, we may observe in practice the operation of 
his critical principles. In the north-western parts of his maps, there are 
very few anachronisms, such as Alvion o: Albion = Great Britain, 
borroved from Pytheas (yet notice the present Alban = Scotland), or 
the presence of a "Rhenish Swabia", dating from Caesar's times. In 
southern Sarmatia, Ptolemy's main prototype was a map, closely con- 
nected with the corresponding source of Pliny, and with abundance of 
antiquated Herodotian names. But Ptolemy has eliminated them all, 
except one, the tribe-name of Bodinoi. The same prototype was the 
first known document which correctly described the Caspian Sea as an 
inland water, and not as a gulf of the northern ocean. And this tremen- 
dous improvement on our geographical ideas was bequeathed to posterity 

^) Cf. the Editio Romana of 1478. — J. Fischer, "An Important PtoL MS.", p. 227. 


through the sole medium of Ptolemy^). On the Tabula Peutingeriana 
from the 4th century, again the old wrong scheme prevails. 

So far Ptolemy's scheme deserves all praise. 

But now we turn to his weak points which cannot escape notice. 

Ptolemy may have been aware of his predecessor's low power of 
topographical and philological divination, but he himself was unable to 
introduce sufficient emendations. He could not discover the wrong inter- 
pretation of physical outlines, nor the regular presence of fancy duplicates 
or triplicates in most parts of Germany, Sarmatia, and Dacia. And 
even where Ptolemy actually improved the maps he did not follow a 
definite principle. It is probable that he scratched out antiquated names 
on the western and southern maps of Marinus, — e. g. it is almost cer- 
tain that the southern part of Sarmatia Europaea with its multitude of 
Herodotian spectres recorded by Mela and Pliny was expurgated by 
Ptolemy in this manner. But why, then, did he not subject the northern 
part of Sarmatia to the same wholesome process of purgation? He has 
there tolerated a long series of those antiquated Herodotian names which 
were conscientiously eliminated in the regions directly contiguous with 
the Roman Empire. It is almost inconceivable that he should have been 
unable to recognize this piece of Herodotian geography, banished by 
Marinus to the Baltic shores but belonging in reality to the shores of 
the Black Sea. And one of the names concerned, Hippopodes = 
''Horsefoot-men", obviously betrays its fabulous nature. In other words, 
the whole mass is a most conspicuous sample of that "mixed stuff" 
which ought to be excluded, according to Ptolemy's own principles. 
Thus he cannot quite escape the suspicion of falsification: he seems to 
have tolerated the "mixed stuff" simply in order to fill out a peripheral 
area of which he really knew nothing. And if that is the case, Ptolemy 
may have proceded similarly when he had to accept or reject the fancy 
duplicates and triplicates delivered by Marinus: he may have regarded 
the despised barbarian names as good enough to be used two or three 
times over in the philological bed of Procrustes, simply in order to fill 
out unsightly bare spots. 

The scheme of Marinus, as delivered by Ptolemy, at any rate remained 
the most terrible chaos. The Ptolemaic maps of northern Europe and 
Asia have, to a great extent, become completely useless, as long as the 
chaos remains unexplored. 

On such grounds, Miillenhoff in his "Deutsche Altertumskunde" III, 
p. 95 etc. denounces Marinus and Ptolemy emphatically, calling them 
"schlimmer als Poeten und Prunkredner", or the "Sudelkoche" of ancient 

') Mullenhoflf, "Deutsche Altertumskunde", 11, p. 95. 

§ 5- Ptolemy's successors 15 

The verdict is no doubt too hard. For, as we saw above, the bad 
qualities do not prevail in all parts of Ptolemy's atlas. And the arbi- 
trary scheme of constructing maps re-appears in most other geographies 
of that kind down to modern times. But at any rate, Miihlenhoff's ver- 
dict marks the culmination of classical geography in an impressive way. 
And the Ptolemaic faults have more or less completely spoiled the modern 
maps of classical Germania down to the year 1914. 


After the time of Ptolemy, a continuous cartographic tradition can 
be traced, represented first by the Tabula Peutingeriana in the 4th cen- 
tury, the local insignia of the Notitia Dignitatum in the 5th, and the 
mosaic map from Madaba in the 6th. These documents are highly 
valuable in order to investigate the development of the Ptolemaic tech- 
nique in several points. The general tendency of their development, 
however, is not an advance, but rather a retrogression. 

The Tabula Peutingeriana, — our most famous relic of classical 
cartography after Ptolemy's atlas, — is a so-called "Itinerary". That 
is to say, it is a mere register of road-distances, meant for wrapping up 
and transporting in a traveller's bag, and therefore it has an extremely 
oblong shape which quite distorts the geographical forms, introducing 
''overlapping" or "telescoping". We may compare it with modern sche- 
matic railway-maps. Its constructor most likely would have been able 
to design a fairly good map of the world on Ptolemaic, lines, — it 
only lay outside his intention to do so. This cartographer was again 
followed by numerous copyists and imitators; they soon surpassed him 
in arranging the whole world artistically according to their private ima- 
gination, but at the same time they lost the ability of constructing more 
accurate maps. Even if some of the same persons mechanically copied 
the Ptolemaic originals, it did not occur to their minds to continue on 
the lines indicated by such superior models. 

In the same measure, as the art of exact cartography declined, the 
tendency towards introducing pictorial and phantastic elements increased, 
finally reducing cartography almost to a mere child's play. Cf. our 
article in "The Scott. Geogr. Mag.", June 19 14. 

Only the reproductions of Ptolemy's atlas remained free from the 
invasion of picturesque barbarism. At the same time, Ptolemy's mediaeval 
copyists were free from critical ambition, contenting themselves with 
mechanical copying. It was reserved to the editors during the humanistic 
age, and to "critical" cartographers as late as 1914, to continue on the 
lines of Marinus — Ptolemy in the sense that they increased the confusion, 
instead of revealing and reducing it. 

i6 Ptolemy's maps of northern Europe 

It was fortunate that the mediaeval copyists so piously and modestly 
respected the character of the original atlas, avoiding alike picturesque 
fancies and would-be-critical emendations. Thus, throughout the middle 
ages, the classical work remained a traditional sanctuary, and it was 
handed down to us through manuscripts, the best of which contain hardly 
any deteriorations worth speaking of. 

Taking it as a whole, we may say that these MSS. represent the 
Ptolemaic atlas in pure, undisturbed condition. The confusion, appearing 
on their maps, is only that which is due to the classical cartographers. 

In the following paragraphs, we shall examine the various forms of 
this confusion. 


In order to penetrate the Ptolemaic labyrinth we will begin with 
examining the different classes of prevailing misreadings or misconcep- 

The Greek constructor of the Ptolemaic atlas was not always successful 
in interpreting his Latin prototypes. His knowledge of Latin appears to 
have been rather inadequate. 

Hermann Miiller has revealed one really classical case^). Ptolemy's 
list of Germanic towns begins with "Fleum, Siatutanda" on the Frisian 
coast. The exact position of Siatutanda is 'defined thus: 29°, 20' of 
longitude, 540, 20' of latitude. The name Siatutanda sounds trustworthily 
"barbarian", at any rate unlike Latin. Still the whole is simply con- 
structed from a passage in Tacitus' "Annals", IV, 72. It is here stated 
that the Roman general Olennius, fighting with some Frisian rebels, re- 
tires to the castle of Flevum. Then in 73 follows the sentence: "Apro- 
nius . . . exercitum . . . Frisiis intulit . . ., ad sua tutanda digressis re- 
bellibus". "Apronius led the army against the Frisians, after the rebels 
had retired in order to protect their homesteads". — Now the riddle is 
solved: "Siatutanda" = "sua tutanda" = "protect their homesteads". 
A slight misreading, — a ^ read as an i — , and a wonderful barbarian 
place-name was ready, worthy of being fixed and defined on the scientific 
map with astronomical data and the rest of it. The town "Protect-their- 
homesteads" still decorates Spruner's "Atlas antiquus" of 185 1, and 
learned critics earnestly discuss the possibility of its continued existence, 
— e. g. Ledebur re-discovers it in Utende at the river Sate^). 

^) "Marken des Vaterlandes" I, p. 114. 

^) "Die Bructerer" p. 180. Both C. MuUer, ed. of Ptolemy, I, 1, p. 266, and Berger, 
"Gesch. d. Erdkunde der Griechen" III, p. 156, are sceptical against H. MuUer's explanation 
but our supplementary observations will show that this is superfluous. 


Other instances of misread Latin words or constructions have also 
been observed. 

Marobudon, town near the Markomanoi, seems to be a castle of the 
Marcomannian king Marbod, originating from the "Annals" of Tacitus, 
II, 62: "Catualda, profugus olim vi Marobodui .... fines Marcomanorum 
ingreditur . . . irrumpit regiam castellumque juxta situm". "Catualda, 
who had ,been expatriated by the force of Marbod, attacks the frontier 
of the Marcom'ans and assails the king's palace and the neighbouring 
castle". See Miiller's ed., I, I, p. 273. 

The detection of these instances of manufactured geography leads us 
to expect more of the same nature, of which the following cases are 

Agrippinensis 11^ IX, 2 in Version I = Cologne is Colonia Agrip- 
pinensis, named after the Empress Agrippina who was born in the town. 
The Greek cartographer did not know that Agrippinensis is a mere 
adjective and therefore dropped Colonia, — that is to say: ignored 
exactly that half of the name which survives till our times. 

In eastern Germany, Ptolemy has the following names of towns, 
written continuously in the context and also placed close by eachother 
on the map: 

Bunition Virunon Virition Rugion Skurgon 

390 30', 55MO' 40030', 55^ 4I^ 54° 30' 42^30', 55Mo' 43°, 55° 

The forms Bunition and Virition might, perhaps, with some difficulty be 
explained as true Gothonic names. But in western Germany we observe 
a town called "Munition", — obviously a Latin "munitio", i. e. a Ro- 
man"fortress". And as* Ptolemy often mutilates initial letters in the most 
unscrupulous way — this will be shown in § 7 — we can no more 
doubt that the "town" Bunition is the same "munitio" in Ptolemaic dis- 
guise, as C. Miiller has already suggested. Most likely, it is again found 
in the third of the above-mentioned would-be-barbarian names, "Virition". 
The remaining three names would then most likely represent tribal de- 
nominations, to be connected with the fortresses. Bunition Virunon, read 
*Munition *Virun6n, would be the "Fortress of the Virunoi", a tribe, 
mentioned by Ptolemy. Virition might be connected either with Rugion 
or with Skurgon: *Munition *Rugi6n = "fortress of the Rugians", or 
''Munition *Skir6n" = "fortress of the Skires". 

The tribe-names Teutonoaroi Virunoi have puzzled scholars greatly. 
Miillenhoff in his "Deutsche Altertumskunde" II, 287, assumed that the 
monstrous form Teutonoaroi must be an arbitrary invention by a Roman 
geographer. But it is simply a Ptolemaic misunderstanding of a Latin 
correction. The prototype had the names written thus: 


i8 Ptolemy's maps of northern Europe 



The letters "vari" simply meant an emendation, intending to re-establish 
the correct reading Varini instead of the arbitrarily distorted form Viruni, 
cf. § 20, g. Ptolemy however regarded *'Vari" as the termination of 
^'Teuton(i)", and thus the monstrous form Teutonoaroi resulted. It was 
built like secondary Latin forms such as Pictavarii, Andegavariij Breonarii 
instead of Pictavi, Andegavi, Breuni, = the modern Poitiers, Angers, 
Brenner. — The fact that "Vari" could be interpreted as the ending 
of "Teuton" shows, that the prototype was written in Latin. 

In Poland, there appears a Ptolemaic tribe with the Latin name 
*Transmontanoi (Codd. & atlas: Tranomontanoi). Miillenhoff, ^'Deutsche 
Altertumskunde" II, p. 84, identifies these people with the Transjugitani, 
mentioned by Ammianus Marcellinus XVII, 12, 12, and signifying some 
Dacian tribe north of the Carpathian mountains. But the Transmontanoi 
evidently belong to the tribal name of Koistobokoi, contrasting their 
northern branch with those Koistobokoi whom Ptolemy's map of Dacia 
places south of the mountains. Our cartographer here again did not 
notice the attributive meaning of a Latin word: out of the "*Coistoboci 
*transmontani", he constructed two separate tribes, the "Koistobokoi" 
and the "Tranomontanoi". (In the same way, he separated the "Ba- 
starnai" from their alter-ego "Peukinoi".) The misreading o for s of 
course contributed greatly to this erroneous statement. 


It is extraordinary that a learned geographer, and a Roman citizen, 
could be so liable to misreading words written in the language of the 
Romans. But that he did so is undeniable, and this fact gives us a 
measure to judge how unscrupulously the Piol. constructor must have 
treated barbarian narhes. 

The conclusion is as evident, as it is important to our valuation of 
Ptolemy's orthography. Still nobody seems as yet to have made this ob- 
servation. The philologists — it is true — sometimes distort Ptolemy's 
spellings in a scarcely less Procrustean manner, than the ancient carto- 
grapher did himself The great linguist and ethnologist Zeuss e. g. alters 
"Daukiones" into "*Skandiones", "Rutikleioi" into "*Turkileioi" = the 
Turcilingi of the 5th century. He correspondingly alters "Veltai" into 
"*Letuai" =: Lithuanians, and out of the Scandian "Leuonoi" Mullenhoff 
forms *"Kyenones" r= "Quaenes" (cf. § 27). The Lithuanians and Quaenes 
are otherwise not mentioned in antiquity; nevertheless, Miillenhofif goes 
so far that he writes "Kyenones" in the Ptolemaic text of his "Ger- 
mania antiqua", without warning the reader that it is a mere conjecture! 



But the same philologists who venture upon such bold conjectures, 
often maintain that Ptolemy has in other points adhered rigidly to the 
original orthography. Whereas he is freely permitted to mutilate middle 
and final sounds, it is a general presumption that he has respected the 
initials in a way which might almost be called reverential. E. g., scarcely 
a single critic hesitates to amend "Busakteroi" into "Burakteroi", or 
"Kognoi" into "Kotinoi", — for here the initial sound is not affected. 
But many dare not with Zeuss correct the Jutlandic "Fundusioi" into 
"*Eudusioi" = "Kudoses" (Tacitus), and the etymology of "fund" gives 
rise to various speculations; Reichard connects it with the island of 
Funen, whereas another explains "Fundusioi" as a nick-name of the 
Eudoses: "Foundlings" instead of "genuine children"! Miillenhoff spends 
a whole portion of learned criticism on proving that Safarik is mistaken 
when interpreting the Sarmatian "Stauanoi" as a distortion of "*Slauanoi" 
= Slavs. 

This distinction between the primary and secondary place is mere 
fancy. It might have been justified, if Ptolemy — or his predecessor 
Marinus — had had the same philological training as his learned critics. 
But the same geographer who read well-known Latin words like the 
most ignorant of grammar-school pupils, would be hopelessly doomed to 
bewilderment, when faced with barbarian forms with which he was for 
the most part totally unacquainted. To him it was no matter of sounds 
or phonetics, — the barbarian names were letters only, — letters without 
sense and interest — , and the beginning was not a bit more protected 
against mutilation than the middle or the end. — If anything rather less. 

To ignore this essential observation makes the treatment of the Pto- 
lemaic orthography completely planless. 

In order to demonstrate the corruption of Ptolemy's initial spellings 
we will instance some examples from Gaul. 

The Gallic names, from Ptolemy's point of view, were barbarian, in 
as much as they were neither Latin nor Greek. But, as Gaul belonged 
to the Roman Empire, nothing could be easier than to ascertain the 
orthography of important names from that province. Nevertheless, Pto- 
lemy's spelling of such names is often most cruelly distorted. Cf. the 
following list: 

Ptolemy Classical Orthography 

Patribatioi Atrebates 

Samnitai ' Namnetai (Ptol.), Namnetes 

Romandyes Viromandui 

Uessones Suessiones 

Subanektoi Silvanectes 

Dueona Devona. Divona 

Modern French Form 

Artois, Arras 




Senlis (metathesis for *Selnis) 




French is known as one of those languages, in which ancient words 
have been radically altered and mutilated. Still, it will be seen at the 
first glance that the modern French forms of the above names are ge- 
nerally much more to be trusted than the would-be-classical spelling in 
Ptolemy's Geography. This observation sheds valuable light on the 
situation within Gothonic regions. We are entitled to expect the full 
analogy here, and we are dispensed of any reverence which would before- 
hand seem due to the "classical" authority. 

The same observation is to be made regarding several of those 
authorities which we must use in order to verify Ptolemy's orthography. 
The works of Strabo and Tacitus often distort the Gothonic names in 
exactly the same cruel manner; in their case, however, the distortions 
may be due rather to the copyists than to the authors themselves. 


To a great extent, the present Ptolemaic orthography of exotic bar- 
barian names must be regarded simply as a field of ruins. 

If, therefore, we examine each name separately, it would in many 
cases lead to nothing. Our chief key of identification must be a survey 
of the entire milieu. 

If we take a whole series of names instead of the single ones, there 
is a certain amount of hope that we may solve the riddles. A skilful 
Procrustes may distort single names into complete obscurity, but he will 
rarely be able to do the same with an entire complexus of them, if he 
does not at the same time disturb their mutual order. 

Let us for example take two of the above-mentioned questionable 
forms, Daukioness and Fundusioi. 

Are we to follow Zeuss who upsets Daukiones into *Skandiones? 
Are we to defend the initial spelling fund? 

The isolated criticism leads to no sure answer. But when we regard 
the entire milieu, things will look quite otherwise. 

Among the Gothons, we know of nearly a hundred sufficiently 
verified tribe-names. Within this number, the initial sound Da occurs 
only once^); the same is the case with the termination dus: the nearest 
assonances, apart from the Eudoses, are the Helisii and the Hellusii 
(Tacitus). Now the only verified name on Da occupies exactly the place 
of the Daukiones, — it is the well known name of Danes. And the 
only verified name on -dus points strongly towards the' neighbourhood 
of the Fundusioi: it is the tribe of Edusii (Eudures) = Eudoses, who 

^) The Dandutoi Ptol. are not verified. 


like the Fundusioi appear jointly with Charudes and Varines (Caesar, 

This verification is decisive. We learn that Daukiones are =: Danes, 
and Fundusioi = Eudoses. It is not simply a suggestion. It is proved 
in the most strict philological sense of the word; otherwise, any 
evidence of combined geographical- linguistical reasoning would be worth 

In the following paragraphs, we shall set forth several collective ob- 
servations, which may assist us in tracing the various distortions of bar- 
barian Ptolemaic names. 


We have mentioned above that different authors assume quite radical 
metatheses: Daukiones <; *Skandiones, Rutiklioi < *Turkilioi, Veltai <Z 
*Letuai (Zeuss). In all of these cases, the assumed "correct" form is a 
mere conjecture, not exemplified in classical times. In the case of 
Daukiones, the milieu undoubtedly proves that the conjecture is wrong. 
The same would be the case with the other assumed unnatural meta- 
theses, but it would be a waste of time to show this. 

There are many cases, however, where the assumption of metathesis 
is natural or necessary. We shall now register some of the most con- 
spicuous cases. 

Form with metathesis Form without metathesis 

1. Dueona, II, VII, 9 Deuona II, XI, 14 

2. Atuakuton II, IX, 5, Version I Atuatokon II, IX, 5, Version II 

(& Context) (Mediolan. Ambros. & Urb. 83) 

3. Asbikurgion II, XI, 5 mountain, Askiburgion II, XI, 10, mountain. 

Version II (Laur. Plut., Med. Version I 

Ambr., Burney) 

4. Bikurgion II, XI, 14, "town" Askiburgion II, XI, 10, mountain 

5. Uispoi II, XI, 6 Usipi, Usipii 

6. Kalukones II, XI, 10 Kathylkoi Strabo VII, p. 291 (i.e. 

*Kaukloi, "smaller Chauks") 

7. Fabiranon II, XI, 12 Foro Adriani Tab. Peuting. 

8. Robodunon II, XI, 15, Version I Eburodunon BEGZ (Eburodanon 

2*^!^, Reburodunon X) 

9. Daros II, XV, i Dravus, the river Drau 

10. Frugundiones III, V, 20 Burguntes II, XI, 8, Burgundiones 


11. Reukanaloi III, V, 10 Roxolanoi III, V, 10 



Form with metathesis Form without metathesis 

12. Boruskoi III, V, lO^) Roboskoi VI, XIV, 9 (in Scythia) 

13. Mysaris III, V, 2^) Tamyrake III, V, 2 (Tamyrakis 

Strabo VII, III, p. 19) 

14. Erkabon III, V, 13^) Sarbakon III, V, 15 

15. Ratakensioi (Racatenses Ed. Ulm.) Rakatai II, XI, 11 

HI, VIII, 3^) 

16. Potula(tensioi) III, VIII, 3 Paloda (or Polonda) III, VIII, 4. 

The metathesis appears frequently, where there is a ^ or 6^ in the 
name concerned. Cf. the following cases: 2. tok ^ kut. 3. kib > bik. 
6. *ukl > ulk > luk. 10. urg > rug. 11. ksolan > kanal. 12. bo- 
rusk > robosk. 13. rak > *kar >> sar. 14. bak > kab. 15. kat 
> tak. 

The inferior MSS. contain several more metatheses, e. g. Maktiadon 
H0W instead of Mattiakon. 

Mil Her suggests that Lakiburgion on the Baltic coast might be a 
distortion of the Rhenish name Askiburgion, but we are not able to 
discover a prototype to which we might ascribe this Baltic duplicate 
(or rather triplicate; the third copy of the name would be Askalingion). 
R. Much suggests the metathesis Melibokos II, XI, 5 > Melokabos II, 
XI, 14. 


A frequent case of distortion is the loss of an initial letter or syllable 
which misfortune may easily happen to barbarian names. In Ptolemy's 
Geography, we notice the following cases, originating from Gaul, Ger- 
many, or Sarmatia. 

1. Romandyes II, IX, 6 

2. Uessones II, IX, 6 

3. Metakon II, IX, 3 (Version II) 

4. Bikurgion II, XI, 14 

5. Setvia II, XI, 14 (Version II) 

6. R(i)usiava II, XI, 14 

7. Robodunon II, XI, 15 (Version I) 

8. Chesinos III, V, i 

Viromandui (in Vermandois) 

Suessiones (near Soissons) 


Askiburgion II, XI, 5 (Askiburgion 

Version II, see § 9). 
Artekvia II, XI, 14 (Version I) 
Biriciana (suggested by C. Miiller 

p. 274) 
Eburodunon BEGZ (Eburodanon 

i:0W, Reburodunon X) 
Acesinus Pliny IV, 83 (sugg. by 


*) Suggested by C. Mtiller. 

§ 10. APOCOPE 23 

9. Mysaris III, V, 2 Tamyrake III, V, 2 (sugg. by 


10. Sturnoi III, V, 10 Basternai III, V, 7 

11. Exobygitai III, V, 10 Hamaxobioi Skythai III, V, 7. 

12. Erkabon III, V, 13 Sarbakon III, V, 15 

In the case of Pagyritai III, V, 10, and Pasyris (*Pakyris), the apocope 
was ah-eady found in the prototype, cf. Pliny Pacyris IV, 84. In the 
case of Agaros potamos III, V, 4 = Sinus Saggarus IV, 82, it is Pliny 
who has erroneously added an initial 5. 

Numerous additional cases of apocope occur in the inferior MSS., 
especially H0W, e. g. Auxones = Saxones, ladua = Viadua, Istulas — 
Vistulas, Ubanektoi = Subanektoi (Silvanecti), Erusioi = Nerusioi 
(Nervii), etc. 

The apocope of 5 in Suessiones was most likely due to a misunder- 
standing of the system of ethnic signs before the names of tribes. On 
the original map used by Marinus or Ptolemy, the ethnic sign before 
Suessiones had disappeared, and consequently the initial 5 was regarded 
as ethnic sign. The result was the present form d Uessones. Cf. our 
article in "The Scott. Geogr. Mag." Febr. 19 14, p. 59. 

We have noticed the following cases. 

1. Romorinon II, IX, i gen. plur. of Morinoi II, IX, 4 

2. Patribatioi II, IX, 4 (Version I & Atribatioi II, IX, 4 

Mediol. Ambr.) 

3. Lugoi Didunoi II, XI, 10 "town" Lugi-Dunon 11, XI, 13 

4. Fabiranon II, XI, 12 F. Abiranon = Foro Adriani Tab. 


5. Pasiakes potamos III, V, 4 Axiakes potamos III, V, 14 

6. Setuako-ton II, XI, 15 Septemiaci VII on the Tab. Peuting. 

7. Teutonoaroi II, IX, 9 *Teutoni-Varini = Teutones 

Auarpoi 11, IX, 5 

8. Ouisburgioi II, XI, 10 Osi Burii^) 

, 9. Exobygitai III, V, 10 Hamaxobioi Skythai III, V, 7 

In no. I & 2, the addition most likely was due to a misinterpreted 
ethnic sign before the names concerned. 

The MS. atlases of Version I write -i- Morinoi, whereas the duplicate 
name is written c^ Romorinoi. We suppose that the R originates from 

^) See Ludw. Schmidt, "Historische Vierteljahrschrift" 1902, p. 80. 


from the sign Y, whereas the following o originates from the point to 

the right of this sign. 

The ethnic sign of Patribatioi is |-^, which may have been misread 
for a Latin P. 

In no. 3, Didunoi, the letters di are simply a misreading of the Greek 
article 'oi. 

No. 6, Pasiakes potamos = Axiakes potamos is = P. Asiakes i. e. 
Potamos Axiakes, "the river A." 

The addition ton in Setuakoton is caused by the Latin figure VII 
added after *Septimiako, cf. Septemiaci VII Tab. Peut. 

The remaining four cases are additions of two separate names. Corre- 
spondingly, numerous MS. atlases (such as the Urbinas 82) write Pro- 
toisidones, originating from the expression of the context "protoi Sidones", 
i. e. "first the Sidones"- 

All context MSS. except Vatican 191 and the best representatives of 
Version II (Laur. Plut., Mediol. Ambros, Constantinop.) write Terakatriai, 
originating from 61 re 'Paxargim xal ol Paxdrai, "both the Rakatriai and 
the Rakatai". 


The distortion prevailing in Ptolemy's barbarian names is in many 
cases of merely accidental nature. But in some cases, we observe the 
working of a general factor, the tendency towards "amending" the un- 
known forms after better known models. 

The tendency generally has a centripetal direction, resulting in a so- 
called "nostrification". That is to say: the names from the far periphery 
are remodelled after those which occur within the Roman Empire, 
especially those from Italy or its neighbourhood. But sometimes it also 
occurs, that a name from the Empire is remodelled after a barbarian 
one from the far north; we might call this a "centrifugal disguise". 

It is only the nostrification which plays a practical role. We may 
now give a list of the cases observed by us. 

T,i. 1 J 1,- The model, after which the t> 1 r 

Ptolemy s spelhng t, V j- • , j Real form 

^ ^ ^ name has been disguished 

I. Samnitai Gaul II, 8, 6 Samnitai Italy III, I, 58 Namnetai II, VII, 8 

and island of Samnis 
near Britany, Pliny IV, 
/ 2. Samnitai Scythia VI, Samnitai Italy III, I, 58 Chainides V, IX, 17 
XIV, ao 
5. RomandyesGaulII,IX,6 Romani.? Italy Viromandui 



Ptolemy's spelling 

4. NerusioiBelgiumII,IX,6 

5. Virunon Germany II, 

XI, 12 
Virunoi ibd. II, XI, 10 

6. Kalukones Germany II, 

XI, 10 

7. Lugidunon Germany II, 

XI, 13 

8. Karrodunon Vindelikia 

11, XII, 3 

9. Pataouion Pannonia II, 

XIV, 4. 
0. AlaunoiSarmatialll, V, 
7; Scythia 

The model, after which the 
name has been disguished 

Nerusioi Italy III, I, 37 
Virunon Noricum II, 

XIII, 3 
Virunon Noricum II, 

XIII, 3 
Kalukones Rhsetia II, 

XII, 2 
Lugodunon Belgium (Ley- 
den) II, IX, I 
Lugdunon Gaul (Lyon) 

II, XI, 12 
Karrodunon Bohemia II, 

XI, 14 
Pannonia sup. II, XIV, 4 
Patauion (Platouion) Italy 

(Padova) III, I, 26 
Alaunoi Noricum II, XIII, 


Real form 


Varini Tacitus 

^Kauklones, cf. *Kaul- 
koi, StraboVII, 291 
*Lugoi Dunoi II, XI, 

Parrodunum (inscr.) 

Poetovio (Pettau) 

The centrifugal tendency appears more or less distinctly in the fol- 
lowing cases. 

Ptolemy's spelling 

11. Semnones Italy III, I, 

1 2 . Sudinoi Germany II, XI , 

1 1 (Sudenoi ADM^") 

Model form 

Semnones Germany (re- 
nowned tribe) II, XI, 
8 & 10 
Sudinoi Sarmatialll, V, 9 
(tribe in Sudauen where 
the Roman merchants 
used to buy amber) 

Real form 

Senones Gaul II, 
VIII, 9 

Sudeta ore, Germany 
(mountain) II, XI, 
5 & II 

It is worth noticing that the nostrification Virunoi instead of Varinoi 
occurred already on an original map, used by the Ptol. constructor. It 
had been corrected by the addition of the letters *'Vari" above "Viru-". 
Cf. § 6. 

As the nostrification introduces in most cases forms from Italy or the 
Alpine districts, and betrays no corresponding inclination towards Greece, 
we may suppose that the Pre-Ptolemaic origin is the general rule. But 
the question cannot be settled without an examination of Ptolemy's entire 
work which we cannot undertake here. 

26 Ptolemy's maps of northern Europe 


It is shown above that a contributor to the Ptolemaic atlas, even if 
only mechanically, strived to identify barbarian names with well known 
ones from the Roman Empire. But it hardly ever occurred to his mind 
to take the trouble of examining whether barbarian names re-appearing 
on the different original maps signify identical or separate entities. 

The phonetic identity of the forms may be noticed by Ptolemy, as 
in the case of "Marionis" and "Marionis No. 2" {^^Magmvlg sreQa") II, 
XI, 12, but this is a solitary exception. 

As soon as the forms are not litterally identical, he registers them 
as different names. The mass of such repetitions have already been ob- 
served by C. Miiller, Chad wick, and Novotny. But it has not yet 
been pointed out how thorough-going the phenomenon is. 

In some cases, the arbitrarily repated names seem to appear thrice 
or even four times. E. g. : 

1. Rakatriai II, XI, 11, Rakatai ibd., Ratakensioi II, VIII, 3. 

2. Nauaroi with town Nauaron (Sarmatia Europaea) III, V, 12 & 13, 
Sauaroi (Sarm. Eur.) Ill, V, 10, town Nauaris (Sarm. Asiat.) V, IX, 16. 

3. Virunoi (Ouirunoi) II, XI, 9, town Virunon (separated from Virunoi) 
II, XI, 12, Auarpoi II, XI, 9, Auarinoi III, V, 8. 

4. Kognoi II, XI, 10, Batinoi ibd., Kytnoi II, XIV, 2, Kotensioi (Kon- 
tekoi Athos Atlas) III, VIII, 3. 

5. Buroi II, XI, 10, (Vis)burgioi II XI, 10, Kuriones II, XI, 11, Bur- 
giones III, V, 8. 

The phenomenon of the repetitions is of capital importance when we 
try to reconstruct Ptolemy's prototypes. We must calculate, therefore, 
how far the repetitions may be authentic or arbitrary. ^ 

In the actually existing nomenclature, repetitions of names are of 
course by no means excluded. Ptolemy himself relates several authentic 
repetitions, e. g. of Brukteroi, Kauchoi, Sueboi, Lugoi, Kampoi, Koisto- 
bokoi, Mediolanion. 

The assumable reliability qf Ptolemaic repetitions may moreover be 
advocated by the fact, that his predecessor Marinus had been extra- 
ordinarily diligent in collecting material, cf. Ptolemy's Book I, ch. VI, 
cited in our § 3. 

In a series of cases such Ptolemaic details which stand isolated 
within the whole of antiquity, are confirmed by mediaeval or modern 
evidences, e. g. Galindai = Galinditae, Kalisia — Kalisz, Marnamanis = 
Marna, Korkontoi = Krkonosce hory, Rakatai = Rakousy, Budoris = 
Biiderich, Vidros = Wetter, Stereontion = Strinz, Amisia = Ems (town), 
Tarodunon = Zarten (mediaev. Zartuna). 


Such observations must of course warn us against categorically dis- 
trusting any non-verified repetitions in Ptolemy's work. Yet they are, on 
the other hand, not sufficient to serve as a categoric guarantee. 

We ought to remember Chadwick's sound critical warning against 
blindly trusting the classical tradition concerning peripheral regions*). 
As we have seen above, the distortion of peripheral names is rather the 
rule than the exception, and this observation is not limited to Ptolemy 
but concerns also other classical geographers such as Strabo and Tacitus. 

A "hapax legomenon" from the periphery of the classical horizon is 
of very low value, — we might be tempted to say: generally worth 
nothing. Concerning such cases, we may set forth the following general 
rule : an identification with another name — even if only possible through ' 
violent emendation — is preferable to the assumption of two separate 
'*hapax legomena". 

In order to obtain plausible results, we may strive to identify the " 
"hapax legomena" with well known names from the regions concerned. 
For the exemplified names from the periphery represent as a rule exactly 
the most prominent ranks, and therefore it is the due right of the well 
known "upper ten" in these regions to claim any neighbouring "hapax 
legomena", if the resemblance is only halfway. 

So much about the occurrence of repetitions generally. The next 
thing is to examine the Ptolemaic cases in particular. 

We mentioned above, that Ptolemy has several undoubtedly verified 
repetitions. If we examine these more exactly, we observe that they are, 
as a rule, designated by differentiating marks ; the Brukteroi and Kauchoi 
are divided into the "greater" and "smaller" ; the Sueboi are divided into 
the Laggobardoi, Aggeiloi, Semnones; the Lugoi into Omanoi, Dunoi, 
Buroi; the Kampoi into Adrabai and Parmai; the Koistobokoi south of 
the Carpathian mountains are contrasted with the Koistobokoi *trans- 

When the verified repetitions, consequently, are often distinguished 
by differentiating marks, most instances without such marks must be- 
forehand be suspected. And as soon as two entire "milieus" of dupli- 
cates appear in fairly corresponding order, their separate existence in 
Ptolemy's geography is evidently due to a cartographer's fancy. 

This impression will be supported if we examine the distribution of 
details statistically. 

It is easy to show that geographic and phonetic unreliability prevails 
in certain parts of the atlas. 

Any observer who regards the reproduction of the Athos Atlas, or 
the reconstructed maps in Miiller's edition or in Erckert's "Wanderungen 

^) "The Origin of the English Nation". 

28 Ptolemy's maps of northern Europe 

und Siedlungen" , will see at the first glance that the distribution of de- 
tails — tribes and towns — is roughly speaking homogenous all over 
the area of Germany. 

This scheme no doubt gives a beautiful impression of all-embracing 
knowledge. But the impression — alas! — is false. Ptolemy's scheme 
of distributing details must, as a matter of fact, be characterized as more 
or less artificial. The Roman ideas of Germany were far from being com- 
plete or accurate. Great parts of the country — especially north-east of 
the Elbe — were almost completely unknown. It is mainly at such 
places that Ptolemy fills out the lacunae by means of duplicates and 
misread Latin words. 

But even the more well-known regions did not escape this sort of 
"making geography". 

In south-western and middle Germany, for example, we find numerous 
tribes with most extraordinary names, never heard of anywhere else: 
Karitnoi, Intuergoi, Nertereanai, Dandutoi, etc. We cannot regard this 
material as a piece of trustworthy local geography, drawn from the 
archives of Roman governors or municipalities. We must suppose that 
the monstrous forms are duplicates of well-known names, — only so 
cruelly distorted that we can scarcely recognize them. 

Thus, taking it as a whole, the abundance of Ptolemaic details must 
be greatly reduced; in the majority of his Germanic and N. W. Sarmatian 
regions the existence of duplicates must be regarded as almost normal, 
so far as tribes are concerned, and there are also numbers of town du- 
plicates. It must only be noticed that the alter-ego of a doubled name 
is sometimes not to be found within the map concerned of the atlas, 
but on a preceding or following one, sometimes in quite distant regions. 
After eliminating the presumably arbitrary duplicates, there still remain 
a considerable number of town-names, peculiar to Ptolemy. But, as con- 
cerns names of tribes, the reduction of his "individual abundance" is in 
many regions practically annihilating. 

It may be convenient to register what remains of Ptolemy's individual 
tribe-names from Germany, Scandia, and the Cimbric Peninsula, when 
the unreliable ones are subtracted. (In the following synopsis, those 
marked with a + re-appear in Latin or Greek literature after Ptolemy's 
I. Germany south of the mountains. Adrabai Kampoi, Parmai Kampoi, 

Rakatai, Korkontoi, Turonoi+ = Teuriochaimai. 
II. North-western Germany. (None). 

III. North-eastern Germany. Siliggai+. 

IV. Scandia. Chaideinoi, Firaisoi-}-, Leuonoi, Goutai-f, Daukiones-f. 

V. Cimbric Chersonese. Saxones-j-, Sigulones, Sabaliggioi, Kobandoi, 


At the same time, we may add some few supplementary "hapax 
legomena" from other classical authors. 

Tacitus (''Germania"). F^osi (N. W. Germany), Lemovii (N. E. Germ.), 
Sitones (Scandia), Reudigni, Auiones, Uithones (Cimbr. Chersonese?). 
Notitia Dignitatum. Brisigavi (S.W. Germ.), Falchovarii (N.W.Germ.). 
Ammianus Marcellinus. Bucinobantes (S. W. Germ.). 

Almost all of these tribes have left some trace of their existence, be 
it in mediaeval tradition, be it in modern place-names. Only the fol- 
lowing have not yet been identified: Adrabai, Parmai, Chaloi, Kobandoi, 
Leuonoi, Sitones. The Sigulones, Reudigni, Auiones re-appear in Widsith; 
the Chaideinoi are the later well-known Heinir in Hedemarken; the Fal- 
chovarii and Brisigavi are inhabitants of Veluwe^), resp. Breisgau, etc. 

Within Gothonic territory, the island of Scandia and the Cimbric 
Chersonese contain the highest number of individual Ptolemaic tribe- 
names, viz. together some lO, against 6 or 7 known from other autho- 
rities. On the other hand, the same regions contribute the smallest share 
to the series of arbitrary duplicates, viz. 2 against 1 5 non-doubled names. 
The duplicates concerned are: Finnoi, with an alter-ego in Sarmatia, and 
Charudes = Farodinoi in Germany. As the alter-egos do not enter the 
Scandian or Cimbric ground, the two districts are completely free from 

The lowest number of individual Ptolemaic tribe-names appears within 
North Germany (apart from the Nordalbingian region, which is regarded 
by Ptolemy as belonging to the Cimbric Chersonese). Along the Ger- 
man coast east and west of Holstein, Ptolemy does not add a single 
tribe-name to the number known from previous or contemporary autho- 

In Sarmatia Europa^a, two thirds of the names along the northern coast 
are transplanted thither from southern regions. 

If half of Ptolemy's Germanic details are proved to be duplicates or 
triplicates, it will of course be a severe disillusionment to those who be- 
lieved in his "abundance". But, on the other hand, it is exactly these 
arbitrary repetitions which make it possible to reconstruct his lost pro- 
totypes. Thus, the gain will be greater than the loss. 


At the same time as the Ptol, constructor creates two or three names 
out of the single ones, he wrongly identifies numerous separate geogra- 
phical details. Sometimes, the identification is due to the presence of 

') O. Bremer, Ethnographie. 



identically sounding names, but it is not always the case. The misplace- 
ments of prototypes are to a great extent due to this sort of misinterpre- 
tation, as we shall see later on. Whereas we shall comment upon the 
cases of false identification separately, when describing the single proto- 
types, we may here provisionally undertake a classification according 
to the geographical categories concerned. 

Categories of exchanged 

district & district 




















Explanation of the mistake. (The letters in brackets 
signify the prototypes concerned) 

district Germania in Belgium (C) mistaken for 
the Germania Megale i. e. Germany (Aa). § 24. 

-^ town Kondate near the inferior Loire, now 
Rennes (C) mist, for Kondate on the middle 
Loire (A). § 24. 

towns Flenio & Matilone {C = Tab. Peuting.), 
mist, for Fleum & Marionis (A). § 24. 

town F(oro) Adriani {C 
for Fabaria (A). § 24. 

Tab. Peuting.), mist. 

river Amisias, an afflux of the Lahn {Ad), mist, 
for the Amisias, debouching into the North Sea 
(Aa). The modern name of both rivers is Ems. 
$ 21. 

fortification & river 

tribe Tungroi in Belgium (C = A), mist, for the 
Tenk(t)eroi in Germany (Aa). § 24. 

the north-western part of the Limes Trans- 
rhenanus (Ad), mistaken for the river Vidros = 
I. Wied & 2. Wetter (Ad), which is again mis- 
taken for the river Ijssel or Vechte (^^). § 21. 

fortification & mountain the northern part of the Limes, and the Miimling 

line (Ad), mistaken for the mountain Abnoba (A). 
§ 21. 

fortification & mountain the eastern part of Limes (Ad), mistaken for the 
. mountain Sudeta (^ or ^i). § 21. 



river & mountain 

the middle & upper course of the Neckar (Ad), 
identified with the western outlines of the moun- 
tains Abnoba & Albia (A), § 21. 

river Rhine (C), mistaken for the mountain Ab- 
noba (A). § 24. 


Categories of exchanged 

river & mountain 

frontier & mountain 

frontier & river 

road & river 

mountain & tribe 

mountain & town 

town & mountain 

coast & mountam 

coast & river 

Explanation of the mistake. (The letters in brackets 
signify the prototypes concerned) 

the inferior course of the Danube, with affluents 
{Ae), mistaken for the Transsylvanian mountains 

(Ac). § 22. 

the north-western frontier of Raetia (Al^), mis- 
taken for the south-eastern outhne of the moun- 
tain Albia (A). § 21. 

the western frontier of the Belgian Germania [C), 
mistaken for the river Rhine (Aa) =. the western 
frontier of Germania megale. § 24. 

the route connecting the upper and inferior 
Vistula (Bi), mistaken for the Vistula itself (A). 
§ 23. 

the mountain Sudeta (^i), changed into the tribe 
Sudenoi (B2). § 23. 

the mountain Asbikurgion (^i), changed into 
the town Bikurgion {B2). § 23. 

the town *Arlaunon (C, now Arlon) localised near 
the mountain Taunus (A). § 21. 

the coast of the Maeotian Sea (£), mistaken for 
the mountains of interior Sarmatia (F). § 26. 

the coast of the Venedikos kolpos, i. e. the 
Baltic (B), mistaken for the river Vistulas {A & 
F). § 26. 

The reader will perhaps at the first sight ask incredulously, how we 
are able to guess at the different sorts of topographic misconceptions pre- 
vailing in the Ptol. constructor's method of working, — they may often 
seem quite impossible to trace. Here again we must answer that the 
entire milieu is the key to the correct interpretation. In order to discover 
the original position of misplaced Ptolemaic details, we must direct our 
attention towards those marked physical features which happen to be in 
the neighbourhood, — either coasts, mountains, or rivers. If a fairly 
correct localisation is effected, when we give the line concerned a new 
name, we may take it for granted that we have discovered the design of 
the original prototype. E. g., we may consider the tribes Ombrones, 
Auarinoi, Frugundiones, Sulones, Finnoi along the river Vistula in south- 


western Sarmatia. These tribes are absolutely unknown in any historical 
or geographical records of the region concerned, and it is at first sight 
clear that the Finns can not possibly be placed south of the Wends on 
the frontier of Prussia and Poland ! But as soon as we replace the Vistula 
by the coast of the Baltic, we obtain a quite correct list of localisations 
which is to be rewritten thus: Ambrones, Ouarinoi, Burgundiones, Gutones, 
Finnoi. Another illustrative case is the Ptolemaic localisation of Me- 
diolanion, Teuderion, Nouaision, Vargiones east of the Rhine; this absurd 
piece of topography will be amended in a satisfactory manner, as soon 
as the mountain Abnoba is replaced by the Rhine. 


In addition to the list of errors, we may make some observations con- 
cerning Ptolemaic features which belong to the category of theoretical 
arrangements. Some of them are arbitrary or directly wrong, whereas 
others may be better founded, but they at any rate point towards a 
collective editorial scheme, and some of them may be referred directly to 
Ptolemy himself 

A collective feature of the atlas is its tendency in favour of schematic 
divisions, and the preference given to the number 3. 

Three classes of regions are distinguished: I indicating the pertinence 
of the towns to the various tribal districts; II with towns, but no ethnic 
classification; III without towns. The distribution of the classes is more 
or less arbitrary. All large islands, except Great Britain, are excluded 
from class I, even if they belong to the very best known radius, such 
as Corsica and Sardinia. The entire Germany is placed within class II, 
although no towns were really known by the Romans between the middle 
Elbe and the Oder. On the other hand, the Cimbric Chersonese is 
placed in class III, although it was decidedly better known that the last- 
mentioned German region. We suppose that the classification is due to 
Ptolemy himself. 

Three classes of towns are distinguished: I with towers, and with a 
a cross as astronomic mark^); II with battlements; III without towers or 
battlements; the astronomic mark in II and III is a point. Class I con- 
tains the towns which are used by Ptolemy as bases of observations 
concerning the length of the midsummerday. Such an astronomic point 
of view certainly betrays Ptolemy as author. 

Three times three islets appear, symmetrically arranged round the 
Cimbric Chersonese: 3 western, 3 northern, 3 eastern. The two versions 

^) Observed by J, Fischer, "Die handschriftliche Ueberlieferung", p. 227. 


of the atlas differ in the arrangement, as I has m ''' m, whereas II has 
= ^=^)- This artistic arrangement can not possibly have occurred on 
the local map from which the design was originally drawn: for a map, 
designed directly on the basis of the Roman marine discoveries in the 
year 5 A. D. would certainly not have indulged in such fancy schemes 
of merely ornamental nature. 

A conspicuous feature of the Ptolemaic atlas is the strong inclination 
of several northern coast-lines towards the north-east, appearing especially 
on the British islands and the Cimbric Chersonese. It may originate 
from the Ptol. constructor, but it may also have occurred on an original 
map, used by him, as it is traditional in Greek geographical literature. 

A third arbitrary arrangement within the Ptolemaic atlas is the limi- 
tation of Germany. The country is represented roughly speaking in a 
square form. It includes the corner between the middle Rhine and upper 
Danube, — a district which had at Ptolemy's times been a Roman do- 
minion for about a century — although one of Ptolemy's sources was a 
special map which represented the Roman frontier wall in the most con- 
spicuous manner. On the other hand, the Cimbric Chersonese and the 
"island of Scandia" are placed apart, within a different statistical class, 
as we have mentioned above; besides, the name "Kimbrike Chersonesos" 
is written on the map with capital letters which rival those of "Ger- 
mania megale". Cf. our § 28. These arbitrary arrangements are evidently 
due to a cartographer whose scheme was more ornamental and geome- 
trical, than truly topographic. 

Prototype A, and perhaps also others of the Ptolemaic sources, con- 
tained the Roman system of roads, or at least the main lines. But such 
details which would have added largely to the practical value of the 
atlas are completely ignored by the Ptol. constructor. 

We have now finished considering the various classes of Ptolemy's 
errors and arbitrary arrangements. In the following paragraphs, we shall 
proceed to the reconstruction of his assumable prototypes. 


The prototypes of Ptolemy's work betray their existence most ob- 
viously in those names which are doubled or tripled. But we may also 
recognize them in those names which appear only once. The fancy re- 

^) Observed by J. Fischer, "An important Ptolemy MS.", p, 229, and "Die handschrift- 
liche Ueberlieferung", p. 229. 



petitions are generally like the backbones in whole bodies or complexes 
still preserving their cohesion inherited from the original prototypes. 

Let us, e. g., take the tribes from the borders of the lower Elbe. 
Firstly, the "Lakkobardoi", i. e. Langobards, appear localised along the 
Elbe directly beside the Saxons. Secondly, their alter ego "Laggobardoi 
Sueboi" appear near the Rhine, directly beside the "Aggeiloi Sueboi", 
or Angles. — Only localisation no. i is correct, whereas no. 2 is due to 
fancy repetition and misplacement. But cohesion with the surroundings 
is disturbed in neither case: localisation no. i correctly shows the Lango- 
bards as neighbours of the Saxons, and no. 2 just as correctly places 
them beside the Angles. Moreover, the cohesion in case no. 2 appears 
at the first glance from the additional "Sueboi", common to both of the 
tribes concerned. 

Similarly, we may in most cases point out whole series of non- 
repeated names accompanying the series of fancy repetitions. In order 
to have a fixed comprehensive denomination, we may unite both cate- 
gories as '^repetition milieus", or, when speaking more definitely, as 
"duplicate milieus" or "triplicate milieus". 

Having stated the existence of such milieus, the next thing is to 
examine from what sort of prototypes they are derived. 

Two main alternatives must be considered. 

Our author — Marinus or Ptolemy — may have read various de- 
scriptions, such as Strabo's "Geography", Pliny's "Natural History", and 
the "German ia" of Tacitus, etc. From these he would have picked up 
the same names three or four times without recognizing their identity, 
and finally he would have tried to distribute the supposed new names 
within the framework of the Imperial Roman map of the world. 

Or, we may suppose that our author did not start from descriptive 
works, but from ready-made maps. Thus, he did not localise every 
supposed new name separately, but reproduced the whole series, found 
on his original maps. 

The first alternative seems to be preferred by Miillenhoff. Cf. espe- 
cially the second volume of his "Deutsche Altertumskunde", wherein he 
deals with the making of Ptolemy's section Sarmatia Europsea.' On 
the map of Germany, there are certainly some cases more or less 
distincly belonging to this category. The most prominent is the famous 
"town" Siatutanda or "Protect-their- homesteads" which has been unveiled 
by Hermann Miiller as an extract from the "Annals" of Tacitus. An- 
other is the town Marobudon, originating equally from the Tacitean 
"Annals". Cf our § 6. 

But generally we are inclined to prefer the second alternative. 

At any rate, it is clear that alternative no. i would make an analysis 
of the Ptolemaic atlas almost hopeless, whereas no. 2 would give a far 


better chance. For the localisations found in the classical descriptions of 
barbarian Europe and N. Asia are very vague and would become com- 
pletely confused when interpreted by a bad philologist such as the Ptol. 
constructor. Whereas a map says more distinctly what it means, no 
matter whether its contents are right or wrong. 

We therefore think that, for argument's sake, we must start from the 
presumption that Ptolemy's atlas has been constructed mainly on the 
foundation of ready-made maps, and not mainly on the foundation of 

Our task will be an attempt to reconstruct the supposed original 
maps or "prototypes" used by Marinus-Ptolemy. The provisional re- 
search, in our opinion, has led to satisfactory results. If the critics will 
not admit it, they may counter-verify our results by undertaking a re- 
construction of Ptolemy's sources on the base of alternative no. i. We 
shall not enter upon this experiment ourselves, — for if alternative no. i 
were really preferable, we should not regard the ultimate results as worth 
the trouble. 

Our paragraphs dealing with the single prototypes will contain the 
following sub-divisions : 

a. Summary of Contents; b. Ptolemaic Localisation; c. Definition of 
Limits; d. General Topographic Scheme; e Statistical Features; f. Oc- 
currence of Duplicates; g. Linguistic Marks; h. Literary Milieu; i. Exa- 
mination of Details; j. Conclusion. 


For the sake of a general survey, we start with a synopsis of the 
Ptolemaic prototypes assumed by us. In this way, their prominent 
features will more easily be realized and compared. Each of the sum- 
maries will be repeated unaltered at the beginning of the paragraph 
dealing with the prototype concerned. — Cf. our figure i which attempts 
to represent the assumable distribution of prototypes. 

A. (§ 18). Collective map describing Europe partially 

or entirely. 
The extension, as specified under i — 5 beneath, would correspond to 
the areas of the local prototypes A, Aa, Ah,, Ac, Ad & Ae, Bi. Pre- 
sumably containing e.g.: i) a physical description of Germany; 2) tribes 
along the German and Cimbric coasts; 3) fortification lines and towns in 
the Roman Limes district between the Rhine and the Danube; 4) tribes 
and towns along the mercantile road from the middle Danube to the 
mouth of the Vistula; 5) tribes and towns in Roman Dacia till beyond 
the Carpathian mountains; but scarcely recording towns in other regions 


36 Ptolemy's maps of northern Europe 

east of the Rhine and north of the Danube. Containing a system ot 
roads. The Latin language is probably used in editing. Originally 
derived from the Imperial Roman map of the world- affinity with the 
Tabula Peutingeriana. 

A. (§ 19). Local map, an oro- and hydrographic description 

of Germany. 
Contains the most detailed representation of German mountains, known 
in classical times; the rivers are represented with less detail. Latin 
language of editing. Affinities with authors of the first century A. D., such 
as Strabo, Mela, Phny, Tacitus. — Serves as main basis of the corre- 
sponding section in A. Cf Fig. 5. 

Aa. (§ 20). Special map; a coast description, stretching from 
about the Rhine to eastern Denmark. 
Including Scania, but not the whole of the Scandinavian Peninsula. 
Detailed observations of headlands and islands; numerous tribes, but few 
or no towns. Duplicates of its names occur in C, D, E & F. Some 
Latin marks. Executed shortly after the expedition of the Roman navy 
to the Cimbric Chersonese 5 A. D. Affinities with Augustus (Monum. 
Ancyr.), Mela, Pliny, less pronounced affinities with Strabo and Tacitus. 
— Correctly amalgamated with A. Cf. Fig. i — 4, 6 — 7, 29. 

Ad. (§ 21). Special map, describing the Roman Limes 
Containing fortification lines, mountains, rivers, and numerous towns, 
but no tribes. No duplicates. Latin marks. Executed after the con- 
struction of the Vallum Hadriani, i. e. towards the middle of the second 
century A. D. Affinity with the Tabula Peutingeriana. — The main 
part is correctly amalgamated with A, but the extremities are extended 
too far towards the north and the south-east. Cf. Fig. 8— 11. 

Ac. (§ 22). Physical map of Dacia. 
Probably with few or no towns. Executed perhaps before the Roman 
conquest. Correctly amalgamated with A. Cf. Fig. 13. 

Ad & Ae. (§ 22). Itineraries describing Dacia. 
Containing rivers, tribes, roads, and towns. Ad and Ae are partially 
duplicates of eachother; scattered duplicates besides occur in Bi, B2 & 
F. Latin marks. Executed after the Roman conquest of Dacia 105 A.D. 
Affinities with the Tabula Peutingeriana (= the Anonymus Ravennas). 
The prototypes seem to have been amalgamated before the times of 
Ptolemy; the map resulting is roughly speaking correctly amalgamated 
with A. Cf. Fig. 12—18. 


Bi & B2, (§ 23). Itineraries, describing the mercantile road 
from the Danube to the mouth of the Vistula. 
Containing mountains, rivers, tribes, a road-line, and towns. Bi and 
B2 are dupHcates of eachother; scattered duplicates occur in Ac & E. 
Latin marks; B2 may have been translated into Greek before the stage 
of Ptolemy. Executed after the introduction of a well established Ro- 
man amber trade under the reign of Nero (54—68 A. D.). Affinities 
with Strabo and Tacitus. Bi is correctly amalgamated with A\ B2 is 
displaced, being introduced directly west of the twin prototype Bi. Cf 
Fig. 19—20. 

C. (§ 24). Itinerary, describing north-western Gaul, Belgium, 
and a part of north-western Germany. 
Containing rivers, tribes and towns. Duplicates occur in Aa and D. 
Latin marks; perhaps translated into Greek before the stage of Ptolemy. 
Close affinity with the Itinerarium Antonini and the Tabula Peutingeriana. 
Displaced towards the east, the Belgian Germania of C being mistaken 
for Germany of A. Cf Fig. 21 — 23. 

D, (§ 25). Local map or description, containing Swabian tribes 

about the lower Elbe. 
Only tribes traceable. A duplicate name occurs in Aa. No Latin 
marks. Affinity with Strabo and especially with Tacitus. Displaced to- 
wards the west, partially from the Elbe to the Rhine. 

E %i F. (§ 26). Collective maps, describing eastern Germany, 
Sarmatia Europaea, Sarmatia Asiatica, and Scythia. 
Containing all sorts of geographical categories; F is besides marked 
by a system of "ethno-topic denomination". E and F are duplicates of 
eachother; scattered duplicates occur in Aa, Ac, Bi, B2. E has Latin 
marks (Sarmatai instead of Skythai F), but seems to have been translated 
into Greek before the stage of Ptolemy. F has only Greek marks. — 
Executed after the introduction of a well established Roman amber trade 
with the Baltic regions during the reign of Nero. Affinity with Pliny, 
including antiquated Herodotian names. — ^ E is placed in eastern Europe 
and northern Asia, not entering Germany; it is turned over, so that 
north becomes west, whereas east becomes north. — F continues the 
eastern parts of A without confusion worth speaking of It is possible 
or likely that F was amalgamated with Sk, before the cornbination of 
the latter with A took place. Cf. Fig. 24 — 26. 

38 Ptolemy's maps of northern Europe 

Sk. (§ 27). Special map or description of the Scandinavian 

Containing tribes only. No duplicates, except Finnoi in E. Greek 
marks. A limited affinity with Tacitus. — Possibly amalgamated with F\ 
finally introduced into the Scanian Peninsula of A (= Aa)\ it is so far 
correctly localised, but compressed within far to narrow an area. Cf. 
Fig. 27. 

a. Summary of Contents. 

The extension, as specified under 1 — 5 beneath, would correspond to 
the areas of the local prototypes A, Aa, Ab, Ac^ Ad & Ae, Bi. Pra- 
sumably containing e. g. : i) a physical description of Germany; 2) tribes 
along the German and Cimbric coasts; 3) fortification lines and towns 
in the Roman Limes district between the Rhine and the Danube; 4) tribes 
and towns along the mercantile road from the. middle Danube to the 
mouth of the Vistula; 5) tribes and towns in Roman Dacia till beyond 
the Carpathian mountains; but scarcely recording towns in other regions 
east of the Rhine and north of the Danube. Containing a system of 
roads. The Latin language is probably used in editing. Originally 
derived from the Imperial Roman map of the world; affinity with the 
Tabula Peutingeriana. 

It may at the outset be taken for granted that the work of Marinus 
was no mere mosaique of local maps or descriptions, freshly amalgamated 
by him, but that it started from more or less collective bases, and one 
of these would have been our assumable prototype A. We are not able 
to investigate it throughout Europe, as it would lead too far. For 
argument's sake, however, it is necessary to point out its possible 
traces within our particular sphere of concern, viz. middle Europe and 

We may here anticipate from the heading "literary milieu" that there 
actually existed a collective map with an extension fairly corresponding 
to that of our Prot. A beyond the Rhine and the Danube. It is the 
Tabula Peutingeriana which contains: A) northern German tribes as far 
east as towards the Elbe, e. g.- Chrepstini = Cherusci; B) towns of the 
Roman Limes between the Rhine and the Danube; C) towns in Roman 
Dacia right north to the Carpathian mountains; D) the tribe of Buri, 
perhaps representing an originally more detailed description of the mer- 
cantile road from the Danube to the mouth of the Vistula. The pre- 
sence of such a collective map is a fact which will remain unshaken, 
even if we do not succeed in proving the existence of a corresponding 


document by means of internal observations from the Ptol. atlas. Con- 
sequently, we may regard the Tab. Peuting. as the main basis for as- 
suming a collective prototype A. 

b. Ptolemaic Localisation. 

Two sharply contrasting strata appear within the Ptolemaic atlas: the 
correctly and the badly localised prototypes. 

The Ptol. maps of Germany and surroundings betray the existence of 
the following local prototypes which are in complete or partial harmony 
with the collective framework of the atlas: 

A = physical map of Germany; An = Denmark and north-western 
Germany (partially corresponding to region A of the Tab. Peuting.); 
Ad = south-western Germany (= region B^ Tab. Peuting.); Ac, Ad & 
Ae = Jazygia and Dacia (= region C, Tab. Peuting.); Bi = the mer- 
cantile road from the Danube to the mouth of the Vistula (= the some- 
what questionable section D of the Tab. Peuting.). — - A, Aa, Ac, and 
Bi are localised correctly. The main parts of Ab and Ad have been 
treated equally. But the northern extremity of Ab invades Aa, whereas 
the south-western seems to be turned the wrong way. Ad and Ae have 
suffered various displacements, although they are not entirely banished 
from their due localisations. 

The following prototypes have all been totally misplaced by the Ptol. 
constructor : 

B2, a duplicate oi Bi\ C =^ Belgium & north-western Germany; 
D = the group of northern Swabians; E = a. collective map of north- 
eastern Germany, Sarmatia Europaea, Sarmatia Asiatica. 

The collective prototype F, a duplicate of E, is on the contrary cor- 
rectly amalgamated with the Ptol. map of middle Europe. 

The local prototype Sk, i. e. the map of the Scandinavian Peninsula, 
is connected with the design of Scania on the Ptol. map. The localisa- 
tion is so far correct, but the scales of the two maps are obviously 
unequal and the Ptol. constructor has not been aware of this essential 
difference (cf. p. 40). 

The distinction of what is correctly and incorrectly localised may 
sometimes be a matter of dispute, but the general fact can scarcely be 
contested that two such strata exist within the Ptol. maps of Germany 
and surroundings. 

It seems to us that these two strata must betray the working of at 
least two different editors. The carthographer who interpreted a whole 
series of local maps fairly speaking correctly, apart from smaller mis- 
takes, would not at the same time be found guilty of misplacing another 
series in the most absurd manner. Our argument is supported by the 
fact that the series of the correctly localised prototypes re- appear ge- 


nerally on the Tabula Peutingeriana, partially with traces of the same 
moderate errors, whereas the Tabula contains no trace whatever of the 
larger Ptolemaic misplacements (those represented by the localisation of 
the prototypes B2, C, D, and E). This observation will be discussed 
more particularly under the heading "literary milieu". 

Consequently, we assign to Prot. A, as a rule, the more or less cor- 
rectly localised prototypes. We except, however, Prot. F and Sk. The 
possibility is perhaps not excluded that Prot. A Sl F should be regarded 
as representing in one stratum a relatively correct map of the world. 
But F, at any rate, possessed an individuality of its own. It appears 
from different observations, viz.: the system of "ethno- topic denomination", 
cf. under d. : the occurrence of duplicates, cf. under e.; the pure Greek 
orthography, cf. under f. Prot. Sk, as we mentioned above, represents a 
scale largely differing from that of A and also the pure Greek ortho- 
graphy of Sk points towards a separate individuality. Most likely, Sk 
had been introduced into F, before the Ptol. constructor amalgamated 
this prototype with A. 

c. Definition of Limits. 

After stating generally the different qualities of the two Ptolemaic 
strata, our next task is to examine in detail how far the superior one 
stretches towards the north-east, — so far we may extend the assumable 
Prot. A, and no longer. 

Along the coast, the extension is easy to define. The superior de- 
sign embraces the German North Sea coast, the Cimbric Chersonese and 
the ''island of Scandia". This area, corresponding to the local prototype 
Aa, stretches far east on the northern side of the Baltic. But, on the 
southern side, the superior design suddenly stops when the base of the 
Cimbric Chersonese is reached: the German and Sarmatian coast of the 
Baltic is a smooth theoretical Hne with no observation of local details 
except the fact that the coast curves towards the north-east when the 
mouth of the Vistula is passed. 

The "island of Scandia", i. e. the peninsula of Scania, must have 
been completely blank, apart from its own name. The seven Scandian 
tribes on Ptolemy's map, including Norwegians and Fins, can not origin- 
ally have been compressed within such a narrow space. Scandia is only 
the fifth part of the Cimbric Chersonese which affords room for practi- 
cally the same number of tribes, (8). As a matter of fact, most of the 
MS. atlases give up the attempt at writing out the names of thfe Scan- 
dian tribes, because the space is insufficient. 

The above-mentioned Baltic coast of Germany with the smooth the- 
oretical outline is almost quite as bare of detail, containing, as it seems, 
only the following verified tribe-names which may be assigned to A: 


Semnones, *Varinoi, Teutones. The rivers Chalusos and Svebos are 
duplicates of the Oder and Vistula, introduced from the displaced Prot. 
B2 by the Ptol. constructor. The frontier of this practically blank region 
is formed by the middle Elbe, the mountain Askiburgion, and the river 

Then follows a better known region stretching from the Oder till 
beyond the Vistula. It is the area of the mercantale road from the 
Danube to the amber coast. The larger part of the Ptolemaic river 
*' Vistula" is simply the line of this road in disguise as it appeared in 
Prot. A (= local Prot. Bi). 

East of the Vistula, the assumable traces of A again disappear. 
Ptolemy decorates tho coast with four rivers, Chronos, Rudon, Turuntes, 
and Chesinos. Three of them at least certainly belong to the misplaced 
Prot. £, being transplanted from the coast of the Black Sea where Pliny 
knows of the rivers Rhode and Acesinus. 

In the inland region towards the south east, we may trace Prot. A 
throughout the map of Dacia which contains traces of relatively correct 
physical observations. We are not able to decide the eventual extension 
of A farther east (cf. under b., p. 40). 

d. General Topographic Scheme. 
When we claim for Prot. A the correct physical framework, the idea 
of accuracy is of course to be understood ''cum grano salis". Numerous 
details of Europe which may belong to A are obviously wrong; e. g. 
the peninsulas of Istria and Chalkidike and the north-westward turning of 
the Rhine are ignored. But it must not be forgotten that in several 
cases Prot. A may have suffered deterioration at the hands of the Ptol. 
constructor, cf. § 15. 

e. Statistical Features. 

Prot. A seems to have contained categories which were eliminated by 
the Ptol. constructor. 

The Roman fortification wall between the Rhine and the Danube was 
represented, NB supposed that the local prototype Ad belonged to the 
elements of A. Our presumption is supported by the fact that a part 
of the wall re-appears on the Tab. Peutingeriana, only mistaken for the 
upper course of the Danube. 

A road system is also indicated. One of its routes is traceable from 
the Sarmatian (= small Carpathian) mountains to the inferior Vistula. 
It is the well-known amber road which has by the Ptol. constructor been 
disguised as frontier-line between Germania and Sarmatia; besides, a 
section of it is erroneously identified with the upper Vistula which flows 
in reality much farther east, apart from the very short initial branch. — 


A whole series of roads are traceable in Dacia, belonging to the local 
prototypes Ad and Ae. The situation of the Ptolemaic towns corresponds 
so exactly to the routes of the Tab. Peuting. that we observe clearly 
how the Ptol. constructor must have effaced the road-lines of the 
original map. 

Ethnic signs, connecting the tribes with their respective towns, seem 
to have occurred within the area of Prot. A, because they have given 
rise to distortion of the Ptolemaic orthography in such cases as pAtri- 
batioi, roMorinoi, Wessones, cf. § lo og ii. We cannot, however, 
discern whether they belonged to the collective prototype A or only 
to some of its local elements. The present quite arbitrary distribution 
of the signs throughout all Continents is clearly due to the Ptol. con- 
structor, cf. § 15. 

A negative criterion is the absence or rareness of that peculiarity 
which we call the "ethno-topic denomination", and which has become a 
directly stereotypic mark of the collective Prot. F. Within the western 
area of Ptolemy's atlas, it is so rare that its presence may be regarded 
as merely accidental. We notice e. g. only two instances north of the 
Elbe, viz. Kimbroi & Kimbrike Chersonesos, and Saxones & Saxon 
islands. There are two instances between the Elbe and the Vistula: 
Sveboi & river Svebos, Virunoi & town Virunon. In Dacia, there would 
have been a natural opportunity of introducing some 4 or 5 cases 01 
''ethnic-topic denomination", cf. § 22, but it has not been used. As the 
ethno-topic denominations abound on the neighbouring Sarmatian ground, 
originating from Prot. F, we may regard their rare occurrence in more 
western regions as a sign that the sections concerned have a different 

Apart from the roads which are traceable on various points, it is 
scarcely possible to point out any marked statistical feature which might 
form a means of defining the area of the assumable prototype A. 

In order to realise the absence of outstanding statistical features, 
marking the area of Prot. A, it will finally be adviseable to regard the 
Ptolemaic inequalities, due to local prototypes within the area of Ger- 
many and its environs. 

a. = Prot. Aa. The Cimbric Chersonese and north-western Germany are 
filled with tribes which seem to be correctly localised. On the other 
hand, the Cimbric Chersonese is entirely bare of towns, and in north- 
western Germany, the correctly localised towns are at least rare. (The 
Ptol. constructor may have eliminated some towns from the Cimbric 
Chersonese, according to his arbitrary scheme, but there could scarcely 
have been many from the very beginning.) 


b. = Prot. Ad. The Limes line in the mountains of south-western Ger- 
many has numerous towns, but no verified tribes. 

c. The eastern side of the Rhine valley from Tarodunon to Mattiakon 
(Zarten-Wiesbaden) has neither verified tribes nor towns. It ought to 
have had ten times as many towns, as occur in C (cf. § 21, d.). 

d. = Prot. Bi {= B2). Bohemia and eastern Germany are well furnished 
both with tribes and towns, and this is the case in both duplicate- 
series of a repetition-milieu. 

e. = Prot. F, A long part of the coast directly east of the Vistula is 
occupied by the lonely name of Venedai = Wends. No towns in this 
section of the prototype. 

f. = Prot. E. The extreme easterly part of the European north-coast, 
in return, is filled with an overwhelming mass of displaced tribes, 
tightly compressed. No towns in this section of the prototype. 

g. — Prot. Sk. The island of Scandia contains only tribes. These are 
correctly localised, as regards their mutual positions, but too tightly 

It will strike the observer that each of the types mentioned is cha- 
racterized by distinctly individual features. Whereas such inequalities 
would be effaced within the territory of the Roman Empire, they could 
not disappear in foreign peripheral regions which supplied a less abundant 
mass of cartographic material. The contrasts here persist, thus forming 
a means of pointing out the various local elements which have been suc- 
cessively combined with the framework of the collective prototype. 

f. Occurrence of Duplicates. 

We assume that the duplicate series Ad and Ac belong to the col- 
lective prototype A, because they re-appear on the Tab. Peuting. Burg- 
iones is — Buroi Bi\ cf. BVR Tab. Peuting. 

Otherwise, the names from the area of A only re appear in the dis- 
placed prototypes, and in F. 

Gaul and Belgium. *Namnitai, Ratomagos, *Bagakon, Askiburgion, 
Morinoi, Vaggiones A =■ Namnetai, Ratomagos, Bogadion, Askalingion, 
roMorinoi, Vaggiones C\ the first four names belong to the contents 
of the Tab. Peuting. 

N. Germany. Lakkobardoi A = Laggobardoi D. 

Cimbric Chersonese. Charudes A — Farodinoi (/??). 

E. Germany. The entire series of A — Bi re-appears in B2. 

Baltic coast. Teuton . . Ouirunoi A — Teutones Auarpoi F, Auarinoi E. 

North-eastern Dacia. Karpianoi A = Harpioi with town Harpis F. 


We do not count the two Marionis, as we regard that of Prot. C as 
a distortion of Matilone Tab. Peuting. 

The line of duplicates in A and F stretching from the Baltic to the 

Black Sea roughly corresponds to the western frontier of the "ethno-topic 

denomination". It might be tempting to regard Ouirunon (read: *Ouari- 

non) as an ethno-topic annexe to Auarpoi (read: *Ouarinoi) F. But we 

have seen above that the "^'Ouarinoi of A, connected with *Ouarinon, were 

already within the Latin stage distorted into *Viruni, and then corrected 

back into . ^^. .. And the distortion started from the town *Ouarinon, 
* viruni 

which was "nostrified" after the well-known Roman town Virunum in 
Noricum. Thus it is scarcely possible to assign Ouirunon to Prot. F. 
It would at any rate require that the prototypes A and F had been 
amalgamated at a very early stage, 

A third alternative must be taken into account, namely that the 
duplicates Teuton . . Ouirunoi = Teutones Auarpoi might belong t5 the 
twin prototypes Bi & B2. — Our reason for assigning the said dupli- 
cates to A and F is found in the triple equation: Ouirunoi A — Auarpoi 
F = Anarinoi E, As Prot. E is an obvious duplicate of F, the pre- 
sence of *Ouarinoi in the one seems to involve its presence in the other. 

g. Linguistic Marks. 

Ptolemy's bad orthography in numerous cases reflects his prototypes, 
betraying a contrast between Latin and Greek ones. The assumable 
collective prototype A — or the local prototypes harmonizing with its 
framework — obviously would belong to the Latin set. 

We observe the following types of Latin residuals: non-translated 
Latin words such as Munition; non- transcribed Latin terminations such 
as -us, -um, -/; -o or -on or one (instead of the correct Greek from -dn)\ 
misunderstood Latin correcture in Teutonoaroi-Virunoi ; misreadings 
pointing towards Latin types such as 6Uessones; non transcribed Latin 
spellings -ng, -nk. 

Somewhat less conclusive, but still noteworthy are the following two 
peculiarities : 

Constant spelling -ones with "omikron" (not with "omega"). 

Constant spelling -aou, -eou (not -au, -eu). 

In these two cases, no analogy could be found in a Latin prototype, 
because the Latin alphabet lacks a similar distinction. Still we believe 
that the said orthographic features are residuals pointing towards Latin 

It seems that the Greek transcription with "omikron" was the estab- 
lished rule for such Latin names which had no settled Greek orthography 


of their own. It was quite natural, because the letter "o" was the same 
in both alphabets. But this conventional rule did not harmonize with 
the tendency of the spoken Qreek language. At least in Ptolemy's atlas, 
the eastern maps obviously prefer -ones (with "omega"), and we must 
suppose that the orthography in these parts of the world was mainly 
based on the principles of the Greek language and represents the verna- 
cular phonetic tendencies of this idiom. Hence we draw the conclusion 
that a constant spelling with -ones ("omikron") points to the presence 
of a Latin prototype, from which the letter "o" was mechanically inherited 
instead of introducing the more vernacular Greek orthography with "omega". 

As to the spellings -aou, -eou, etc., it might at the first sight seem 
natural to regard these as indicating a Greek prototype, because no 
distinction between -aou and -a?i, -eou and -eu existed in the Latin 
alphabet. But although the Greeks possess the distinction, lacking in 
the Latin alphabet, they practically do not use it within their own 
"sphere of interest". Whereas the Romans, in spite of the want of distin- 
guishing letters, seem to have actually observed the distinction in their 
spoken language. This again must have been noticed by the Greek car- 
tographer who transcribed the Roman maps in his own language. The pre- 
sence of the distinction, therefore, seems to be a trace of Latin prototypes. 

So much about the Latin marks generally. We shall now regard 
their geographical distribution. 

Britain: (H)orrea, Tarvedum, Verubium, Virvedrum. 

Spain and Portugal: Aistuaria, 2 Lukos, Libunka, Konkana, Segis- 
amonkulon; Lakippo, Baisippo, Akinippo, Oiasso, Asso, Mago, Ursone, 
Sisapone, Alauona. (The correct Greek form is introduced into the 
names of the important mercantile centres: Tarrakon, Barkinon, Oliosip- 
pon = Tarragona, Barcelona, Lissabon.) 

Gaul and Belgium : Agrippinensis (Latin adjective) ; Tungroi, Obrinkas 
(Cod. Vatic. 191); Kessero, Karkaso; Kossion; (the correct Greek form 
in the important name Narbon); Piktones, Senones, ((5)uessones, Redones, 
Vaggiones, Loggones (all with "omikron"; no exceptions). 

Cimbric Chersonese: Misreading Fundusioi for *Eudusioi (Eudoses 
Tacitus); Saxones (beside Sigulones)- 

Germany: appellative "Munition" and the identical Bunition (and 
Uirition, again = Munition.?); termination -one, on(e) in Munition etc., 
Singone, Grauionarion (= Grinarione Tab. Peuting), Fleum, Semanus, 
"town" Lugi-dunon = the tribe of Lugoi Dunoi. ; correcture "vari" above 
Virunoi, mistaken by Ptolemy for Latin plural; Alkimoennis; Tenkeroi 
(Vat. 191), Angrivarioi, Singone, Asanka (and LAKKOBARDOI < 

Pannonia: Saldis (Latin dat. plur.), Akuminkon (two places), Akvinkon. 

Illyria: Kurkum, Oouporum, Stulpi, Ausankalei. 

46 Ptolemy's maps of northern Europe 

Italy: Angulos, Anxana (Vat. 191). 

Sarmatia: *Transmontanoi, Karpianoi (Latin termination); Piengitai. 

Dacia: Salinai, Pirum, Angustia (misreading for Augusta), Sangidaua, 

Moesia: Karsum, Singidunon. 

Egypt: Karkum. 

We have not registered the cases of the spellings -aou, -eou, because 
they are too frequent; e. g. Treoua in N. Germany, Deouona in S. W. 
Germany, Seouakes and Karaouagkas in Noricum, Noouai in Moesia. 

As contrast to the Latin residuals, the Greek ones must be considered. 
We shall name some instances. 
/ Denomination Skythai (instead of the Latin correspondence Sarmatai). 

Greek descriptive words: alsos (in Limios alsos). 

Misreadings, pointing towards Greek types: ^uarinoi, ^uarpoi < 
(9uarinoi, POYriK^IOI < POYriK^IOL 

Constant spelling GG (not NG): Laggobardoi, Aggeiloi. 

Constant spelling -ones (with "omega", not with "omikron"): Gythones. 

Constant spelling -AU, -EU (not -AOU, -EOU): Nauaroi, Sauaroi, 

Regarding the distinction of prototypes, most of these marks are not 
so conclusive .as the Latin ones. For the introduction of Greek lexical 
and orthographic emendations could be undertaken even at the very last 
stage before the issue. Nevertheless, we may suppose that pure domi- 
nation of Greek marks and absence of any Latin residuals will in most 
cases point towards Greek prototypes. 

From this presumption we may except the regions with predominating 
Greek nationality and besides some important mercantile centres with 
traditional Greek orthography. At such places, a Greek editor would 
naturally efface any traces of Latin prototypes. As a matter of fact, the 
toleration of Latin residuals within Greek domains is almost excluded 
(solitary exception: Karkum in Egypt, Codex_ Urbinas 82, noticed by J. 

The result of our observations is that the predominance of the La- 
tinisms agres with the above-mentioned characteristics of Prot. A. We 
stated above that the duplicates Teuton- Ouirinoi Karpianoi — Teutones 
Auarpoi Harpioi mark a line of contact between the prototypes A and 
F, at the same time forming the western frontier of the "ethno-topic 
denomination", peculiar to the latter prototype. Exactly the same con- 
trast appears through the linguistic criteria: on the one side we have 
the Latin .correcture *"vari" above Ouirunoi, and the Latin termination 
in Karpianoi, — on the other we have the Greek misreading Auarpoi 
instead of Ouarinoi. 


Various classes of evidences could scarcely support eachother in a 
more satisfactory manner. 

As we mentioned, it is of course not strictly necessary that all of the 
Latinisms observed must originate from the collective prototype A] 
several might have been introduced from local prototypes. We therefore 
shall repeat the cases concerned, when commenting on those local proto- 
types, which harmonize with the framework of A. But, taking it as a 
whole, it can scarcely be doubted that the Latinisms are a practical 
means of pointing out generally the sphere of Prot. A. 

h. Literary Milieu. 

In order to orientate the reader about the general milieu, we shall 
give two chronological lists. The one contains a series of described 
events, political or mercantile, which influenced the history of geography 
in northern Europe before Ptolemy's times. The other contains the most 
important geographical and historical publications before Ptolemy. We 
include some works from the period after Ptolemy's death, because they 
may reflect his sources. 

List of political and mercantile events. 
58 B. C. Caesar fights the Swabians and other Germans on both 

sides of the Rhine, "Bell. Gall." I, IV, VI etc. 

12 B.C. seq. Drusus and Tiberius begin the occupation of north- 
western Germany. Vellejus II, 97, Dio Cassius LIV, 31. 

c. 2 B. C. King Marbod of Bohemia establishes the great Swabian 

Empire. Strabo VII, 290, Tacitus, "Ann." II, 45. 

B. C. Domitius Ahenobarbus settles a flock of Hermundures 

within a territory left vacant by Marbod's Marcomans. 
Dio LV, 10. Firm mercantile relations between the 
Romans and Hermundures are established, lasting for 
more than a century. Tacitus, "Germania" ch. 41. 

5 A. D. Tiberius camps on the border of the lower Elbe. The 

Roman navy visits the Cimbric Chersonese. Augustus 
"Monum. Ancyr.", Strabo VII, 293, Vellejus II, 106, 
Pliny II, 167. 

9 A. D. The Roman dominion over interior Germany is destroyed. 

Vellejus II, 117, Dio LVI, 18. 

17 A. D. King Marbod's great Swabian Empire breaks down. 

Tacitus, "Ann." II, 44—46. 

48 . Ptolemy's maps of northern Europe 

21 A. D. seq. The Romans repeatedly interfere with the conflicts 

of German tribes in Bohemia and Moravia. Tacitus, 
"Ann." II, 63; XII, 29; "Germ." ch. 42. 

47 A. D. After repeated campaigns in north-western Germany, 

^ the Romans definitely give up the coast between 

the Elbe and the Zuider Sea. Tacitus, "Ann." 
XI, 20. 

Betw. 54 & 68 A. D. A firm mercantile connection with the Prussian 
Amber Coast is established. Pliny XXXVII, 45. 

69 — 70 A. D. Rebellion of the Batavian chief Civilis against 

Rome. Tacitus, "Historiae" IV, 12 seq. 

c. 85 A. D. Masyos, king of the Semnones about the lower Elbe, 

makes a voyage to Rome. Dio LXVII, 5. 

c. 90 A. D. Establishment of the Roman Limes district between 

the middle Rhine and upper Danube. Tacitus, 
"Germ." ch. 36. 

Shortly bef. 98 A. D. The Boructres in north-western Germany are de- 
feated and "almost exterminated" by their neigh- 
bours. Tacitus, "Germ." 36. 

105 A. D. Trajanus conquers the Dacian regions south and 

east of the Carpathian mountains. 

Betw. 1 17 & 138 A. D. Hadrianus completes the fortification wall of the 
Roman Limes district between the Rhine and the 

166—180 A. D. The Romans are engaged in war with the nations 

beyond the middle Danube, such as the Marcomans 
in Bohemia and the Dacians in Poland. Dio LXXI 
seq., Jul. Capitolinus XXII seq. 

List of publications. 
27 — 20 B. C. Agrippa, "Commentarii" ^). 

7 B. C. Map of the world, made by order of Augustus 

(Chorographia Augusti; lost). 

before 14 A. D. Augustus, Monumentum Ancyranum. 

^) Cf. MtiUenhoft", "Deutsche Altertumskunde", III, p. 212 seq. 


c. 1 8 A. D. Strabo, Geographia. 

29 A. D. Vellejus Paterculus, Historia Romana. 

c. 40 or 50 A. D. Pomponius Mela, Chorographia. 

']'] A. D. Plinius, Naturalis Historia. 

97 A. D. Tacitus, Historiae. 

98 A. D. — , Germania. 
c. 115 A. D. — , Annaies. 
c. 211 — 229 A. D. Die Cassius. 

c. 286 — 305 A. D. Julius Capitolinus, Bellum Marcomannicum. 
4th century A. D. Itinerarium Antonini. 
4th — A. D. Tabula Peutingeriana. 

At first sight, it may seem a difficult if not hopeless task to attempt 
to make positive statements concerning the literary milieu of Prot. A. 
For as long as the contents and limits of the prototype are not even 
approximately pointed out, we have no firm base for making literary 

This is true. Nevertheless, we^ may for argument's sake set forth 
some provisional remarks. 

It is natural to suppose that the original foundation of Prot. A was 
the lost Chorographia Augusti, the Imperial Roman map of the world, 
finished in the year 7 B. C. and later no doubt subjected to several 

The Roman horizon towards the north was greatly enlarged through 
the naval explorations along the German and Danish coasts in the year 
5 A. D., and through the contemporary and subsequent undertakings in 
interior Germany, military as well as mercantile. The last important in- 
cidents of this epoch are: the intermeddling of Rome with the affairs of 
Bohemians and Quades, about 21 — 50 A. D., and the establishment of 
a firm mercantile connection with the Prussian amber coast, about 60 A. D. 
The new discoveries were described in the local prototypes A, Aa, and 
Bi, resp. a physical map of Germany, a map of the German and the 
Danish coasts, and a map of the road to the amber coast, cf. §§ 19, 
20, 23. They were indubitably introduced into the framework of the Im- 
perial map of the world, in consequence of its repeated revisions. Corre- 



spending literary milieus are represented by the geographers, Strabo, 
Mela, Pliny, and Tacitus. 

During the reign of Domitianus, (8i — 96), the corner of Germany 
between the middle Rhine and upper Danube was transformed into a 
Roman '^Limes district", and its frontier walls were completed under the 
Emperors Trajanus and Hadrianus the latter of whom reigned since 1 1 5 
A. D. In the year 105, Trajanus conquered the part of Dacia lying be- 
tween the river Theiss and the Black Sea. Through these conquests, the 
Roman Empire obtained its largest extension along the northern side of 
the Danube. The cartographic results were the local prototypes Ab = 
the Limes district, and Ac, Ad 81 Ae = Dacia. 

These maps were also introduced into the framework of the collective 
map. Their main contents were placed correctly, even if several details 
were misinterpreted. 

With the additions mentioned, Prot. A seems to have reached its 
accomplishment. We have assumed above that the relative correct Ptol. 
localisation of the prototypes Ad, Ac, Ad & Ae, etc. was due to another 
cartographer than the one who introduced Prot. 32, C, D, and E in 
the most confused manner. Besides, the linguistic marks of the more or 
less correctly localised prototypes point towards Latin authorship, whereas 
at least two of the displaced prototypes contain Greek marks. 

Whereas the additions to the Augustean horizon are in previous lite- 
rature only reflected by descriptive works, now at last the literary milieu 
supplies a correspondence in cartographic form, viz. the Tabula Peutin- 
geriana. It is a most prominent feature of this document that it contains 
the Roman Limes district and Roman Dacia, thus representing the stand 
of the Empire after the large conquests in the beginning of the second 
century A. D. The existing edition of the Tabula, it is true, introduces 
elements from a somewhat later epoch, — freshly formed German tribal 
names such as Franks and Allemans, and numerous place-names betraying 
the spreading of Roman nationality throughout Dacia; at the same time, 
the entire Cimbric Chersonese and the greater number of details from 
the lost Roman province in northern Germany have been left out, — 
evidently because these regions had long since passed out of Rome's 
practical sphere of interest. Nevertheless, the correspondence with our 
assumed Ptolemaic Prot. A is unmistakable. We also notice that the 
frontier wall of the Limes is traceable on the Tabula, as in A (= Ab\ 
and that the exact correspondence of the Dacian towns in both docu- 
ments betrays that Prot. A contained the same road-system as the 

Supposing that the author of the Tabula extracted Prot. A or a 
closely related map, we should draw attention to a negative fact which 
may perhaps be of some importance to our conclusions. The Tabula 


contains no single trace of displacements corresponding to the Ptol. 
localisation of the prototypes B2, C, D, and E, There is a most inti- 
mate correspondence, it is true, between the Tabula and the displaced 
Ptol. prototype C, but the names concerned on the Tabula all correctly 
hold their place in Belgium, exactly as the corresponding section does 
in the assumed Prot. A\ cf. e. g. the names Namnetes, Ratomagus, 
Bagacum, and Asciburgium, appearing with relatively correct localisation 
in Prot. A and on the Tabula, and with displacement in the Ptol. section 
derived from Prot. C. 

It must of course be admitted that the Tabula leaves out the larger 
part of that area within which the Ptol. displacements occur. Con- 
sequently, the negative evidence is not so valuable as it would have 
been if the area concerned had been copiously represented. Nevertheless, 
there are sufficient regions where displacements of the Ptolemaic sort 
might have been expected: the *Redones from Rennes might have been 
banished to the middle Loire, the *Namnetes from Nantes to the Seine; 
*Langobardi might have occurred at the middle Rhine, *Usipii near the 
Schwarzwald, *Chattuarii at the source of the Danube, etc. In our 
opinion, it is not very likely that these and similar displacements should 
have occurred in the source of the Tabula, and all have happened to be 
eradicated by the author of this map, — quite accidentally. It is a far 
more reasonable alternative to suppose that hardly any such displace- 
ments occurred in the source, extracted by him. There is one exception, 
it is true, but it only confirms the main rule. We have^ stated above 
that the localisation of Prot, Ab, Ad & Ae within our assumable Prot. A 
betrays some errors, e. g. Ad and Ae have been incorrectly combined. 
It is all the more worth noticing that the section Dacia of the Tabula 
contains exactly the same incorrect combination of the two prototypes 

To sum up, we hold that the internal examination of Ptolemy's maps, 
supplemented by the comparison with the Tabula Peutingeriana, seems 
to point towards the existence of a collective prototype A as defined 

The next question is: who was the author? 

One chronological fact is evident: he must have been at work still 
after 115 A. D., in order to introduce the Vallum Hadriani and the 
established system of Roman roads in Dacia. 

The observation would be conclusive as to the autorship, if we as- 
sumed with A. Herrmann^) that the years about icx) A. D. were the 
epoch when Marinus was composing his atlas. Then the author of Yxo\..A 
would simply have been Marinus himself. In this case, the displaced 

') "Zeitschrift des Vereins ftlr Erdkunde zu Berlin", 1915. 



prototypes B2, C, D, and E, would most likely have been introduced by 
his editorial heir, Ptolemy, 

However, we see no strict necessity for placing the working of Ma- 
rinus as early as 100—120 A. D. This date cannot be deduced from 
the Ptolemaic preface which makes no mention of his lifetime. As Ptolemy 
most probably lived to witness the beginning of the Marcomannian war, 
166 A. D., nothing prevents us from placing the work of his predecessor 
about 140. 

Then the author of Prot. A would have been an anonymous carto- 
grapher. He would most likely have been of Roman nationality, as the 
area of Prot. A is so constantly characterized by Latin marks. His 
anonymity cannot surprise us, as we ignore equally the author of the 
Chorographia Augusti from the year 7 B. C. Perhaps, the author of 
Prot. A was only a revisor who introduced the latest acquired local maps 
into the otherwise ready-made collective map. 

The subsequent development would be clear: Marinus would have 
introduced the entire series of displaced maps, such as B2, C, D, and 
E (perhaps also the correctly localised collective map F\ Ptolemy would 
have added nothing, except those few Asiatic and African maps which 
he enumerates in his preface, ch. XVIII. 

Our assumption seems to agree with the literary portraits of Marinus 
and Ptolemy, such as we may draw them on the base of the latter's 
work. Marinus, according to Ptolemy, was a gatherer of material, whose 
energy in collecting was enormous^ but whose power of criticism was 
characterized as insufficient. Such qualities would correspond exactly to 
the uncritical introduction of original maps, with absurd localisation, 
evidently undertaken in order to fill out bare spots. Ptolemy, on the 
other hand, according to his own words, has only contributed little to 
the collection of fresh material. He puts the main stress on the astrono- 
mical fixation of the localities, and on the elimination of antiquated de- 
tails. He has, it is true, tolerated numerous inherited wrong represen- 
tations, and he has not always been sufficiently strict in carrying out his 
own critical principles. But it is easily understood that Ptolemy dared 
not correct his renowned predecessor's maps of peripheral northern re- 
gions which lay far beyond his own horizon. And the partial lack of 
systematic strictness is no sufficient reason for assuming that a critical 
author like Ptolemy would indulge in uncritical heaping of material, 
directly against his own principles. 

i. Examination of Details. 

See the corresponding sections in the §§ dealing with the local pro- 
totypes A, Aa, Ad, Ac, Ad, and Bi, of which A is composed. 


j. Conclusion. 

Owing to the provisional impossibility of examining the entire Ptole- 
maic atlas, our preceding researches consist too much of guess-work. 
Such "pioneering hypotheses" are, however, necessary. And the reader 
need not fear that the guessing will prevail equally in the following 
paragraphs, dealing with the local prototypes: here, the sphere of research 
will be easier overlooked and penetrated. 


a. Summary of Contents. 
Prot. A is an oro- and hydrographic map of Germany. It contains 
the most detailed description of German mountains, known in classical 
times; the rivers are represented with less detail. Latin language of 
editing. Affinities with authors of the first century A. D., such as 
Strabo, Mela, Pliny, Tacitus. Cf. Figures 2, 3, 5 and L. Schmidt (Seeliger's 
"Hist. Vierteljahrschrift" 1902) who has already assumed a corresponding 
prototype. Cf. also the additions in Appendix A (after § 29). 

b. Ptolemaic Localisation. 
The framework of A forms the basis of Ptolemy's map of Germany. 
It is correctly localised, as it must be, for the frontier rivers, Rhine and 
Danube, allowed of no mistake. 

c. Definition of Limits. 

The introduction of other prototypes into the area of A has already 
been described in the paragraph dealing with the collective prototyped, 
§ 18, and need not be repeated here. 

We only emphasize that the rivers Chalusos and Svebos and the 
river-name Viadua belong to the details introduced by the Ptol. con- 
structor, and that the mountain Abnoba A has been displaced towards 
the north-east, owing to the amalgamation with Prot. Ai^: the Ptol. 
Abnoba in reality corresponds to the Vallum Trajani of Prot. Ad, whereas 
its northern extremity may conceal the mountain *Taunus of Prot. A, 
still reflected by the Ptol. position of the town Ar-taunon, cf. under e. 

d. General Topographic Scheme. 
The Ptol. design of German mountains and rivers is highly remark- 
able because of its excellence which may be called almost unrivalled 
throughout the entire atlas. It is e. g. decidedly better than the corre- 
sponding description of Gaul, although the latter country had been a 
Roman province for more than two centuries when Ptolemy was making 


his atlas. Such a physical map of Germany seems to point towards the 
existence of an individual prototype. Its elements, when taken separately, 
might certainly be attributed to the above-mentioned local prototypes, 
such as Aa, Ab, Bi. But the fusion into an excellent physical map of 
Germany seems to point to the authorship of one person, — a topo- 
grapher with very special experience. 

e. Statistical Features. 

Ptolemy records the names of no less than lo mountains or woods 
in Germany. The town Ar-taunon may point towards the original pre- 
sence of a number ii, the Taunus, even if the town itself is in reality 
the Belgian Orolaunum, now Arlon or Aarlen, transplanted by the Ptol. 
constructor from Prot. C. 

There are not so many rivers, as several Ptolemaic ones must be 
eliminated: Vidros belongs to Prot. Ab, Chalusos and Svebos to B2, and 
Viaduas is a duplicate of Vistula. But, on the other hand, the Ptol. 
constructor may have eliminated names of rivers occurring in Prot. A. 
We conclude this from the fact that his map of Germany contains no 
less than three anonymous affluents of the Danube. 

We have not been able to discover any tribes or towns which must 
necessarily have filled out the framework of A. Its contents may have 
been merely physical. 

f. Occurrence of Duplicates. 

Two of the mountains, belonging to the complexe of A, re-appear 
in the duplicate series of Bi & B2, viz. Asbikurgion (alias Askiburgion) 
= Bikurgion, and Sudeta = "tribe" Sudenoi. They must, then, have 
occurred in the special maps describing the mercantile road from the 
Danube to the inferior Vistula. But this assumption does not imply that 
they were omitted in A\ they seem to form absolutely indispensable 
links in the mountain system of the latter prototype. Melibokos A has 
by R. Much been identified with the town Melokabos, belonging to 
Prot. Ab, and we have had the same idea independently; but the dupli- 
cate would in this case most likely have existed in actual nomenclature, 
as Melokabos belongs to a well verified list of Roman frontier fortresses 
(Prot. Ab), and corresponds to the present Miltenberg. 

g. Linguistic Marks. 
Latinism: Semanus, to be supplemented: saltus ("wood"). 

h. Literary Milieu. 
The elaboration of Prot. A is due to the military and mercantile 
undertakings of the Romans during the first century A. D. The eastern 



area, extending from the Danube to the inferior Vistula will be investi- 
gated in the paragraph dealing with Prot. Bi & B2. 

The main features of Prot. A re-appears in the works of all geo- 
graphers from the first half of the century. Cf. the following synopsis. 

Ptolemaic map (supplemented by 


*eninsula Cimbric Chersonese Cimbric Chers. 




a row of islands islands along N. W. 
along N. W. Ger- Germany 

"town Fleum" 

Cimbri in the Co- Cimbric headla: 
dan Gulf, i. e. 
on a peninsula 

islands in the re- 23 islands aloi 
gion of the tide N.W. Germai 
(= North Sea) 


"town" Fabira(non) 


Fabaria = Bi 






Hercynius Hercynius 

Abnova (do. T 


"town" Ar-taunon 

"town" *Teutiburgior 




Askiburgion (do. Bi 

& B2) 
Sudetsi (do. B J &B2) 
Sarmatika ore 


Taunus (do. Tacitus) 

(Teutoburgiensis saltus, Tacitus) 




Albis Albis 

Amisias, Amasias 


Amisis Amisis 




Visurgis Visurgis 
Vistula Vistla, Visculus 

It appears from the comparison that the Ptol. map puts the main 
stress on the orography and in this respect it remains unrivalled. Other 
classical authorities only add little to the Ptolemaic selection, viz. Caesar: 
Bacenis; Tacitus: Silva Caesia; Dio: Vandalika ore; Tab. Peuting. : Silva 

56 Ptolemy's maps of northern Europe 

The Ptol. hydrography of Germany is distinct and good, but it is not 
so unique as the orography of the same section. The Ptol. main rivers 
are known collectively by Mela and Pliny, and, with one exception, also 
by Strabo. And each of these authors, as well as Tacitus, adds rivers 
which do not appear on the Ptol. map, viz. Strabo: Lupias, Salas; Mela: 
Lupia, Flevo, Moenis (and s^yamps Suesia, Metia, Melsyagus); Pliny: 
Flevus, Marus, Duria; Tacitus: Lupia, Nabalia, Adrana, Cusus; cf. also 
Marc Aurehus: Qranua; Ausonius: Nicer. We have, however, mentioned 
that the Ptol. constructor seems to have left out river-names occurring in 
Prot. A, and it is also not excluded that the Ptol. towns Fleum and 
Nabalia were originally accompanied by the homonymous rivers, known 
from Tacitus. 

If we consider the additional material of Strabo, Mela, Pliny, and 
Tacitus, we shall notice that apart from two exceptions, the names con- 
cerned are all quite individual to each of these authors. 

Consequently, a general correspondence between Strabo, Mela, Pliny, 
and Tacitus, only takes place at such points where it is shared by the 
Ptol. map of Germany. 

We regard this fact as a further indication that Prot. A was a docu- 
ment which fundamentally influenced the classical ideas about Germanic 
geography. It furnished the main framework not only of the Ptolemaic 
map, but also of the descriptive representations of the same regions. 

i. Examination of Details. 

Semanus is Fichtelgebirge, the centre of the middle German moun- 
tains. In German, the name may have sounded simply Sema; the 
ending -nus would be a Latin addition. The ancient name seems to be 
preserved in Cechian as Smrciny. 

Sudeta = Bohmerwald. The present localisation^ north-east of Bo- 
hemia is absurd, — a fatal consequence of the superstition that only 
Ptolemy's text and not his atlas must be regarded as conclusive. 

Gabreta = Baierischer Wald, or perhaps some southern extremities 
of the Bohmerwald. . 

Luna, and the Sarmatian mountains, might be respectively the Moravian 
hills and the small Carpathian mountains. But they may perhaps also be 
interpreted thus: sm. Carpathian mountains and Tatra. 

Askiburgion, the "Ash-mountain", is generally identified with the 
Jesenik which means the same in Slavonian. The mountain Jeschken or 
Je§ted in northern Bohemia may perhaps also reflect the ancient name. 

Melibokos is = the Thiiringerwald, according to Ptolemy's map. If 
the same name occurs in Melokabos Ad (read: *Melobakos), now Milten- 
berg, its area would have extended to the western course of the Main, 
including the mountain Speshard. The element -bokos is = "beech", 


occurring in several German names of mountains, such as Deutschbuch, 
Albuch etc. 

Tauno-, in Ptolemy's Ar-taunon, = Feldberg in Hessen. The present 
use of the name Taunus is of course a learned invention. 

Abnoba = the Schwarzwald; the Ptol. displacement of the name has 
been mentioned above under b. Cf. Chr. Mehlis, "Die klassischen 
Namen des Schwarzwaldes", in "Petermanns Mitteilungen", 191 4, p. 74, 
where the extension of the Abnoba is shown by means of Roman in- 

Albia = die rauhe Alb, or Schwabische Alb. 

j. Conclusion. 

The individual existence of Prot. A is in the first line derived from 
the impression which the observer receives from the physical design. The 
general correspondence with the geographers of the first century A. D. 
affords a support, even if it must be admitted that the evidence is 
somewhat meagre, as it is in most cases limited to the category of very 
important names. The Ptolemaic amalgamation of Prot. A with other 
original maps greatly contributes to effacing its prominent qualities, and 
as long as observers contented themselves with the modern Ptolemaic 
maps reconstructed from the text, there would be still less chance of a 
favourable valution. When modern scholars have hitherto unanimously 
placed the Ptolemaic Sudeta north of Bohemia, not south of this country, 
their mistake betrays that they regarded the Ptolemaic design as hope- 
lessly confused. The study of the hitherto despised MS. atlas will here, 
as in other points, contribute to a juster valuation of our assumed Prot. A. 



a. Summary of Contents. 

Prot. Aa is a special map; a coast description, stretching from about 
the Rhine to eastern Denmark; including Scania, but not the whole of 
the Scandinavian Peninsula. It contains detailed observations of headlands 
and islands; numerous tribes, but few or no towns. Duplicates of its 
names occur in C, D, E, and F. Some Latin marks. The prototype 
would have been executed shortly after the expedition of the Roman 
navy to the Cimbric Chersonese 5 A. D. Affinities with Augustus 
(Monum. Ancyr.), Mela, Pliny, less pronounced affinities with Strabo and 
Tacitus. Cf. Figures i — 4, 6 — 7, 29. 

58 Ptolemy's maps of northern Europe 

b. Ptolemaic Localisation. 

In its present appearance, Prot. Aa has been used in order to sup- 
plement the older collective map which originated from the times before 
the Roman discoveries along the North Sea and the Baltic. This carto- 
graphic process of amalgamation was in most cases carried out success- 
fully. Most likely, it was accomplished before the stage of the Ptol. 

c. Definition of Limits. 

The displaced Prot. C, on the Ptolemaic map, invades the German 
part of Aa, covering it all over with towns. The Cimbric Chersonese, 
however, remains completely untouched. 

The western German part of Aa is invaded by Prot. Ab, as the Ptol. 
constructor exaggerates the Limes Transrhenanus and the southern river 
Amisias so far that they reach the North Sea. 

On its southern periphery, Prot. Aa touches the displaced prototypes 
B2 and D, which generally do not invade its area. Prot. D offers the 
most marked contrast. As soon as the German tribes of Prot. Aa stop, 
those of D continue. A sharp line of demarcation is formed by the 
three Swabian tribes of D, stretching from the Rhine to the Oder. Only 
one single tribe of Aa transgresses the line, viz. the *Tenkteroi. The 
sudden appearing of the i>-set shows that the prototype Aa did not go 
farther south than to middle Germany. Then the space was left blank, 
capable of receiving the interpolated Ptolemaic mass of names. 

The extension of Prot. Aa towards the south-east is easy to observe: 
evidently, the southern coast of the Baltic remained unexplored and was 
therefore expressed by a smooth theoretical line betraying no topographic 
experience. The contrast to the relatively detailed design of Scania is 

South of the Baltic, the eastern outposts of Aa touch the north- 
western outposts of F\ Teuton . . Ouirunoi Aa = Teutones Auarpoi F. 

Prot. Sk, i. e. the Scandinavian Peninsula, is amalgamated with the 
blank map of Scania in Aa, perhaps through the intermedium of F, 
cf. §§ 26—28. 

d. General Topographic Scheme. 

The physical design of Prot. Aa is first class. Its coast description 
of north-western Germany is excellent. That of Denmark is simply 
astonishing, ^- for we must take into account that the country remained 
unexplored till the year 5 A. D., and that the Romans had no opportunity 
of continuing their explorations after that date. It is scarcely conceivable 
how the Roman officers could discover so much during some few months' 
stay near the Danish coasts. 


It must be added that the description of the Danish and Saxon 
archipelagoes no doubt suffered deterioration at the hands of the Ptol. 
constructor who introduced an arbitrary artistic arrangement, viz. the 3 
times 3 islets surrounding the Cimbric Chersonese. 

e. Statistical Features. 

Prot. Aa, as we mentioned above, contains mainly tribes, whereas 
Prot. B2 and C contain a copious selection of both tribes and towns. 
Ady on the other hand, contains towns and no tribes. It must, however, 
not be forgotten that the Ptol. constructor may have increased the con- 
trast, by leaving out all details from the Cimbric Chersonese except the 
names of tribes (and of surrounding islets). Cf § 15. 

Within Prot. Aa, we notice some instances of "ethno-topic de- 
nomination", viz. Kimbroi & Kimbrike Chersonesos, Saxones & Saxon 
islets, Virunoi & town Virunon. The occurrence of this feature, however, 
can scarcely be said to constitute a predominant system, such as in 
Prot. F. 

f. Occurrence of Duplicates. 

Chaimai, Kaukoi mikroi, Askiburgion = Kamauoi, Kalukones, Askalin- 
gion C (perhaps also TuUsurgion = Tulifurdon represent a duplicate of 
Aa and C). 

Lakkobardoi, Charudes Aa — Laggobardoi, Farodinoi B. 

Teuton-_, Ouirunoi (Virunoi) Aa — Teutones, Auarpoi F, Auarinoi £. 

Marionis Aa = Marionis C is a pseudo-duplicate, as the name be- 
longing to C seems to be a mutilation of Matilone on the Tabula Peutin- 
geriana. Cf. § 24, f. 

g. Linguistic Marks. 

Latinisms or misreadings pointing towards Latin script. 

Cimbric Chersonese. Fundus'ioi misread for *j5"«dusii. Saxones (with 
"omikron"; versus Sigulones). 

Germany. Bunition = Munition in Ad. Fleum ; Tenkeroi, Angrivarioi ; 
LAKKOBARDOI misread for *LANKO- < *LANCO-; misunderstood 
correction *"vari" above *Viruni; Teuton- (with '^omikron"); Treoua. 
No typically Greek marks. 

h. Literary Milieu. 

Prot. Aa represents the topographical information collected during the 
time of Roman rule over N. W. Germany. The prototype is of some- 
what later origin than the Imperial map of the world, for the former 
was executed by the year 7 B. C, whereas the Roman dominion over 
N. W. Germany did not reach its zenith before 5 A. D. After the 

6o Ptolemy's maps of northern Europe 

downfall of Roman power in the year 9 A. D., and after the Romans 
had in 47 A. D. definitely given up their last positions along the coast 
of N. W. Germany, the topography of these regions soon ceased to be 
generally known, — a fact stated directly by Tacitus, "Ger mania" ch. 41 : 
"Albis .... flumen inclitum et notum olim; nunc tantum auditur". On 
the following pages, we shall show through a series of details that Prot. 
Aa agrees with the authorities from the first century A. D., whereas it 
has marked differences from the stage of Tacitus. As to Strabo, we 
should be led to expect that he would present distinct points of resem- 
blance with Prot. Aa, because he wrote at the beginning of the century. 
But, singularly enough, he rather agrees with the geographers of the 
Tacitean stage. To a great extent, his lack of knowledge is obviously 
due to the fact that he would not believe in the Roman discoveries 
north-east of the Elbe, as he states emphatically VII, p. 294. 

The chief milieu of Prot. Aa is represented by the authors Augustus, 
Mela, and Pliny, as we shall now indicate through a series of ob- 

1. More or less distinct knowledge of numerous islands in the North 
Sea and between the Cimbric Chersonese and Scania is common to Aa, 
Mela, Pliny (III, 6, resp. IV, 96). Strabo at least knew of islands along 
the coast of north-western Germany, whereas he ignored those of the 
Baltic (VII, p. 291). 

2. The name of Scandia is common to Aa and Pliny (IV, 104). 
Pliny's identical name Scadinauia (IV, 96) may be compared with Mela's 
Codanouia (III, 6, 54). 

3. Distinct knowledge of a large gulf behind the Cimbric Chersonese 
is common to Aa, Mela, Pliny (ibd.). 

4. Distinct knowledge of a Cimbric Chersonese forms a prominent 
point of resemblance between Aa, Mela and PHny (III, 3, 32, resp. II, 
167, IV, 96). Strabo also knew of the Chersonese (VII, p. 292), — 
only he would not admit that it was situated north of the Elbe. Tacitus 
had no distinct idea of a Cimbric Chersonese, and at the stage of the 
Tab, Peutingeriana, this idea had disappeared from the horizon of the 

5. Distinct knowledge of the Kimbroi as neighbours of the Charudes 
(= the present Himmerboers beside the Hardboers) betrays a close 
affinity between Aa and Augustus. Cf. Pliny's headland Chactris beside 
the Cimbri (IV, 97); but the name is also spelt Thastris. 

6. The contiguity of the Cimbric Chersonese with the Teuton(oaroi) 
reflects Mela and Pliny who represent the Cimbri and Teutones as neigh- 
bours (III, 3, 32, resp. IV, 99). Both Aa and Mela, like Prot. F, re- 
present the Teutones as a Baltic tribe, although with different localisation 
[Aa and F \ in western Pomerania, or on the island of Riigen; Mela 


(III, 6, 54): on the island of Codanouia, i. e. either Sealand or Scan- 

7. The absence of the Angles on the Cimbric Chersonese (in the 
district of Angel) is common to Aa, Strabo, Mela, and Pliny. This 
negative feature is in contrast to the scheme of Prot. D and Tacitus, 
the only two classical authorities to whom the Angles are known. 

8. The Swabian group does not appear, for the Langobardoi Aa are 
mentioned without the addition of "Sveboi". It is similar to Pliny who 
does not represent the Hermunduri as belonging to the Swabian group. 
Mela, at the best, mentions the Swabians quite by the way^). This 
scheme was a natural consequence of the fact that the great Swabian 
Empire, to which the Langobards belonged, had been ruined in the year 
17 A. D. Strabo here differs from Aa and Pliny on equally natural 
grounds, because he wrote before the catastrophe mentioned and, con- 
sequently, still knew the Langobards as subjects of the Swabian Empire. 
A sharper contrast to Aa and Pliny is offered by Prot. D and Tacitus, 
for here the Swabians are emphasized in spite of their political downfall; 
it is a sort of metachronism which is avoided in the older set of 

9. The sub-division of the Brukteroi is common to Aa and Strabo 
(VII, p. 291)^). The Tabula Peutingeriana represents them as undivided, 
whereas it knows of sub- divisions among the Franks. Tacitus directly 
asserts that the Brukteroi had lately been almost exterminated by their 
neighbours ('^Germ." ch. 33)^). Consequently, we must assume that Prot. 
Aa originates from the times before the said catastrophe. 

10. The Angrivarioi, according to Aa, are placed on the eastern 
side of the Weser, and the Kauchoi only occupy the coast region. Ac- 
cording to Tacitus ("Germ." ch. 33 & 35), the Angrivarii had lately ex- 
tended their territory towards the west, conquering the Bructeri. At the 
same time, the Chauci had advanced in eastern Hannoveria so far, that 
they touched the Chatti, i. e. the inhabitants of Hessen. Here again 
Prot. Aa represents the older stage. 

So much for those authorities whom we may regard as forming the 
main milieu of Prot. Aa. 

It still remains to add some few words concerning the eventual resem- 
blance with the milieu of Prot. Z>, especially with Tacitus. 

^) Mela, III, 5, 45 mentions "Baeti" or "Boti" who are in Pliny's quotation of the same 
passage replaced by "Svebi" (II, 170). 

^) The sub-division of the Chaucs is more generally stated: by Aa^ Strabo (Kaukoi & 
*Kaulkoi) VII, p. 291, Vellejus II, 106, Pliny XVI, 2, Tacitus, "Ann." XI, 19. 

') The words of Tacitus must not be taken quite literally. The Brukteroi were by no 
means exterminated, as they re-appear on the Tab. Peutingeriana, and still as a well-known 
tribe till the 9th century (Bede etc.). 

62 Ptolemy's maps of northern Europe 

The following cases must be taken into account. 

tribe Dulgubnioi = Dulgumnii, Tacitus, "Germania" ch. 34. 

— *Eudusioi = Eudoses, Tacitus, "Germania" ch. 40. 

town Askiburgion = Asciburgium, Tacitus, "Germania" ch. 3, Asciburgio 

Tab. Peuting. (Askalingion Prot. C). 

— Nabalia = river Nabalia, Tacitus, "Historiae" V, 26. 

— Fleum = fortress Flevum, Tacitus, "Annals" IV, 72. 

— Siatutanda(!) = "ad sua tutanda", Tacitus, "Annals" IV, 73. 

Such cases cannot prove that Prot. Aa had the same close affinity 
with Tacitus as with the older geographers. The preserved remnants of 
Mela's and Pliny's works give only fragmentary ideas about the northern 
horizon of these authors. The horizon of Tacitus is much better ex- 
emplified, — we may suppose, that his preserved works illustrate his 
knowledge of Germanic tribes in a fairly exhaustive way. Thus it may 
be a mere accident that Ptolemaic tribes like Dulgubnioi and *Eudusioi 
re-appear only in the works of Tacitus and not in those of Mela, Pliny 
etc. We may add, that even if Mela and Pliny do not mention the 
fortress of Flevum, they know at least the Vlie-stroom, from which it 
has drawn its name (Mela Flevo, Pliny Flevus). The monstrous town 
of "Siatutanda" or "Protect-their-homesteads" is certainly fabricated on 
the base of the Tacitean "Annals", but it does not necessarity imply that 
the blunder was due to the author of Prot. Aa, — the name may just 
as well have been interpolated by a succeding editor. 

The main thing is the fact that the general topographic ideas of 
Prot. Aa harmonize with those of Mela and Pliny, and not with those of 
Tacitus. This fact remains unshaken in spite of the names mentioned 
which re-appear in Tacitean works. 

i. Examination of Details. 

It may be regarded as superfluous to comment upon all physical 
details of Prot. Aa. Their general correspondence with nature is striking, 
whereas nobody will demand of the first map of Denmark ever designed 
that it should be completely free from error. We may content ourselves 
with considering some special points which want explanation. 

The islands of Alokiai have by some scholars been identified with 
the present Halligen along the west coast of Slesvig; so e. g. on the 
map of Germania in R. Kiepert's "Formae Orbis Antiqui", published 
19 14. This identification is improbable from the phonetic point of view, 
and quite impossible from the topographical. The classical form of the 
name would scarcely have begun with a Latin H, resp. a Greek spiritus 
asper, which might easily be dropped. The initial letter would rather have 


been either Ck or K, in Latin C, and even if Ptolemy may drop any 
initial letter, there is no reason for this suspicion here unless the topo- 
graphy would lead us to it. The topography, however, directly excludes 
it, for the Alokiai, on the Ptolemaic map, are not the islands west of Sles- 
vig, but clearly those forming the northern extremity of Jutland, as it 
was already stated a century ago by such Danish scholars as Bredsdorff. 

The present Ptolemaic map certainly exaggerates the distance of the 
Alokiai from the southern shore of the Limfjord, but this representation 
need not belong to Prot. Aa, — it may be a part of the Ptol. con- 
structor's artistic scheme of arranging the islets round the Cimbric Cher- 
sonese. Whereas the insular districts Ty and Vendsyssel north of the 
Limfjord are nowadays connected by an isthmus, the Ptol. map assumes 
the absence of this connection, as it leads 3 channels from the Limfjord 
directly into the bay of Jammer-Bugt. It is possible that the Roman ob- 
servers were mistaken, but at the same time their error would be very 
explicable, for the middle part of the isthmus mentioned consists of hills 
arising to a considerable height within surroundings of low level: such 
a hilly country would like an island when observed from the sea at 
some distance (Bredsdorff). On the other hand, the possibility is by no 
means excluded that the Ptol, map may be right, for the Limfjord has 
changed its western outlet several times, and so it may very well have 
possessed an extra outlet towards the north. The general correctness of 
Prot. Aa speaks in favour of the latter alternative. Provisionally, we, 
must leave the question unsettled, but it is possible that geology may 
in the future give a decisive answer. It has already been suggested, 
without any reference to Ptolemy's map, that channels from the Limfjord 
to the Jammer-Bugt existed about the beginning of our era. If such 
theories proved correct they would thus find their literary verification in 
the classical geography. 

The Ptol. map of Scandia also requires some consideration. If we 
regard the design as given by the Cod. Urbinas 82, or by several other 
MSS., we shall not be particularly struck by its likeness. But as soon 
as we compare the corresponding Mount Athos map, we shall receive a 
different impression, cf. Fig. 27. Here there is an unmistakable individual 
likeness with the actual form of Scania. We notice: the point of Kullen, 
projecting towards the west; then the coast of the Sound with gentle 
inclination towards the south-east; then the south coast, running straight 
west-east; and finally the coast leading towards Bleking with strong 
north-eastward direction. It seems scarcely conceivable that such a 
naturalistic design could be merely accidental. In our opinion, it must 
be derived from the first-hand observations made by the Roman officers 
in the year 5 A. D. 

The names recorded by Prot. Aa are to a great extent preserved till 

64 Ptolemy's maps of northern Europe 

mediaeval or modern times in the local nomenclature. Frisioi = inhabi- 
tants of western Frisia. Their town Fleum, evidently named after the 
Flevus or Vlie Stroom. Chaimai, misplaced by Ptolemy, = inh. of the 
mediaeval Hamaland south-east of the Zuyder Sea. Brukteroi = inh. of 
the mediaeval Borahtra Gau. The Kauchoi in northern Hannoveria seem 
to be the O. E. Hugas whose name survived in the mediaeval Parisian 
district of Hug-merki. Angrivarioi = the mediaeval Angrarii, one of the 
main groups of the Saxons. Lakkobardoi or Langobards = the mediaeval 
Bardi in the present district of Barden-Gau. Virunoi, read *Varinoi = 
the people who lived at the river Warnow in Mecklenburg; the Wendic 
tribe of Varnabi may have been their descendants who had adopted the 
nationality of the Slavs. Saxones = inhabitants of Holstein that was 
in mediaeval tradition designated as "Saxonia antiqua", "Old Saxony". 
Sigulones = the O. E. Sycgas, mentioned in the Widsith poem beside 
the Saxons. Sabaliggioi = the present Sallingboers in Sailing; their 
shire — in Danish Sailing Syssel — in mediaeval times extended farther 
south towards the centre of Jutland. Fundusioi, read: *Eudusioi, neigh- 
bours of the Charudes, are the sEdusii or Eudures mentioned by Caesar 
as fellows of the Harudes on the expedition against Gaul in 58 B. C. 
Tacitus, "Germania" ch. 40, mentions the Eudoses beside the Angles as 
worshippers of the goddess Nerthus. The comparison with Caesar and 
Tacitus shows that Ptolemy is right in placing the tribe among the Jut- 
landers. Charudes = the present Hardboers or Hasselboers in Hard- 
Syssel, a shire in western Jutland. They seem to have moved thither 
during the migration ages, as the Ptol. map places them on the east 
coast. Their ancient localisation may still be reflected by the district 
name of Hadsherred on thie east coast, in mediaeval times Harz Haeret 
(Werlauff). Kimbroi — Himmerboers in Himmerland, the mediaeval Himber 
Syssel. Skandia = Scania, O. N. Skan-ey. 

Among the Jutlandic tribes, we miss the Angles as inhabitants of Angel 
in Slesvig. Their absence, however, cannot surprise us, if we regard the 
fact that the exact observation of coast lines stops at the southern edge 
of the Baltic, The Roman explorers in the year 5 A. D. evidently did 
not land south of the Little Belt. We only hear of negotiations with 
the Kimbroi and Charudes who both lived north of this channel. Even 
if the explorers caught the names of some Mecklenburgers such as the 
Varini, such informations were merely sporadic, and we cannot wonder 
if other names from those vaguely described coasts were ignored, such 
as that of the Angles. 

It is worth noticing that three of the Ptolemaic names of Jutlandic 
tribes are preserved by inhabitants of peninsular districts, viz. Sabaliggioi, 
Charudes, Kimbroi. Peninsular shape of districts always tends towards 
preserving the ancient names, cf. the cases of Kent and Cornwall. 


j. Conclusion. 

Prot. Aa must be called well verified, both from topographic, 
statistical, and literary points of view. It could scarcely be expected to 
betray itself more neatly. 

Whereas the prototype does not especially enlarge our knowledge of 
the region between the Rhine and the Elbe, the description of the 
Cimbric Chersonese is a document of unrivalled importance in the carto- 
graphic history of Denmark. This map, designed during the expedition 
5 A. D., was destined to remain the only map of Denmark worth 
speaking of for almost 1500 years. It was not surpassed till the Dane 
Claudius Clavus designed a map of his country, as it looked in the 1 5th 
century, and even he dared not emancipate himself from the famous 
Ptolemaic scheme^). 


a. Summary of Contents. 

Prot. Ad is a special map, describing the Roman Limes Transrhenanus. 
It contains fortification lines, rivers, and numerous towns, but no tribes; 
no duplicates; Latin marks. The prototype would have been executed 
after the construction of the Vallum Hadriani, i. e. towards the middle 
of the second century A. D. Affinity with the Tabula Peutingeriana. 
Cf. Fig.s I, 2, 4, 8 — II, 30 — 31 , and our article in Paul & Braune's "Beitrage 
zur geschichte der deutschen sprache und literatur", vol. XLI, pag. 17 
seq., where we provisionally discuss the objections of an anonymous 

b. Ptolemaic Localisation. 

On Ptolemy's map, Prot. Ad fills out the entire south-western corner 

of Germany. 

The main part of Ad, i. e. between the upper Rhine and Danube, 
is roughly speaking correctly localised, but the northern and south- 
eastern extremities are misinterpreted or displaced in various ways. 

The mouth of the river Vidros, in the region of the middle Rhine, 
was identified with the mouth of the river Ijssel or Vechte, debouching 
into the Zuyder Sea. Correspondingly, the river Amisias, an affluent of 
the Lahn, was mistaken for its larger name-sake, the present Ems which 
debouches into the North Sea (already suggested by C. Miiller). It was 
obviously the existence of two rivers Amisias which mislead the Ptol. 

*) Cf. A. Bjornbo and C. Petersen, "Der Dane Claudius Clausson Swart", 1909. 


66 Ptolemy's maps of northern Europe 

constructor. And the consequence was that the part concerned of Prot. 
A6 was stretched far too far towards the north. 

Apart from this Procrustean extension, the rivers Vidros and Amisias 
have been subjected to various metamorphoses. 

In reality, the Ptolemaic Vidros represents two rivers: one is the 
present Wied, debouching into the Rhine, and the other is the present 
Wetter, debouching into the Nied, an affluent of the Main. The actual 
courses of these two rivers have disappeared, being replaced by the 
north-western part of the Limes Transrhenanus. 

The river Amisias, according to Ptolemy, starts east of the mountain 
Abnoba which reflects in its northern extremity the wing of the Limes 
in the Wetter district-, and the homonymous town Amisia lies east of the 
Abnoba. In reality, the corresponding river Emisa or Ems starts from 
the north western side of the Limes, where also the hononymous town 
Ems is situated. It is easy to understand that the Ptol. constructor felt 
obliged to "correct" the original map, as soon as he identified the river 
Amisias of the Limes region with its better known name-sake in north- 
western Germany. 

The eastern outline of the Abnoba and the northern outline of the 
Albia reflect the Limes without displacement, but farther east the traces 
of Prot. Ad become less certain. 

It might seem as if the eastern Limes had been absorbed by the 
Ptol. mountain Sudeta = Bohmerwald, but, on the other hand, details 
from the extremities of the Limes perhaps occur farther south. We must 
leave this question for the examination of details. '^ 

c. Definition of Limits. 
Owing to the above-mentioned misconceptions of the Ptol. constructor, 
the northern extremities of Prot. Ad invade the area of prototype Aa. 
On the other hand. Ad is invaded by the contents of the dislocated pro- 
totype C, e. g. Nouaision (i. e. the present Neuss), Vargiones (i. e. 
Vangiones), Uispoi (i. e. Usipi), and Chaituoroi (i. e. Chattuarioi). The 
confusion, however, causes no serious trouble, as the towns and riverg 
belonging to the Limes region are generally easy to point out. 

d. General Topographic Scheme. 
If we subject the Abnoba and Albia to exact examination, using the 
design in the Cod. Urbinas 82, we shall notice that they betray a marked 
difference from other Ptol. mountains. The form of the two chains 
mentioned, especially of the Albia, contains traces of a more minute 
design than we are otherwise accustomed to. And, above all, both 
chains are interlarded with towns, a quite extraordinary feature in the 
Ptolemaic orography, cf. under the heading "statistical features". 


The details constituting Prot. Ad must be spared for the heading h, 
where they will be properly discussed. Provisionally, we may only 
emphasize the observation that the traceable outposts of Ad cling to a 
series of distinctly visible physical lines which are all disguised expres- 
sions of Roman fortifications. The first section is the so-called river 
Vidros, the second is the mountain Abnoba, the third is the mountain 
Albia. Not all towns of Ad, it is true, lie west or south of this com- 
bined line, — there are about half-a-dozen of outsiders, viz. Stereontion, 
Munition, Amisia (with river Amisias), Kanduon, and Grauionarion. But 
almost all of these only lie at a short distance from the demarkation 
line, so that they may be regarded as what the French military language 
calls "entfants perdus", i. e. advanced positions. 

e. Statistical Features. 

Prot. Ad is characterized by the presence of fortification lines and 
numerous towns, whereas tribes are lacking. All of the surrounding pro- 
totypes contain tribes, — so Aa, B2, C, and D. Aa and D have few 
or no towns. 

Also a more intimate statistical examination of Prot. Ab shows its 
marked individuality. 

Along the right border of the middle and upper Rhine, representing 
full two thirds of the entire river-course, Ptolemy has only one single 
town, viz. Tarodunon, the mediaeval Zartuna, now Zarten. We might 
also count Mattiakon, the present Wiesbaden, but the map removes it 
far away from the river. On the other hand, the neighbouring mountains 
Abnoba and Albia are overloaded with towns, amounting to about 14. 
They are literally interlarded with towns, for Kantioibis, Devona, Sego- 
dunon, Lokoriton, and Melokabos are placed inside the mountain strip 
and the two first mentioned have given rise to "lowland cauldrons" on 
the copies designed by Donis, cf Fig. 4. 

Such a distribution, from the statistical point of view, is obviously 
absurd. We should have expected a dozen Rhenish towns for every 
single mountain town, not the opposite proportion. Especially, we miss 
Aurelia Aquensis, now Baden, the capital of the Grand Duchy of the 
same name, and Brisiacus, now Breisach, the capital of the district 
Breisgau. How did it occur to Ptolemy's mind to distribute the population 
in this extraordinary way? 

The explanation is no doubt to be sought in the assumption that 
the original map, used by the Ptol. constructor, did not really describe 
mountains, but another sort of geographical category. If we compare the 
design of mountains and forests in other parts of the atlas, we shall 
certainly find plenty of incisions, — e. g. the forest Gabreta in the New 
York MS. includes no less than four, corresponding to the town vignettes 


68 Ptolemy's maps of northern Europe 

of Eburon, Strevinta, Meliodunon, and Arsikva, cf. Fig. 4. But it rarely 
occurs that the mountains or forests are literally interlarded with towns; 
in the Valencia MS., map of Spain, we observe e. g. two cases, one in 
the Pyrenees, and the other in a south-eastern mountain, see the repro- 
duction in J. Fischer's article, "Iberica" 19 14, p. 105. After noticing this 
fact, it will strike the observer that the mountains Abnoba and Albia 
contain no less th^n half a dozen; we mentioned 5 above, and the MS. 
used by Donis must have added a sixth one, viz. Bomoi Flavioi, for 
here his design shows a corresponding "lowland cauldron". 

It was this observation that first caused us to suggest that the original 
map of the regions concerned must have been a special plan of the Limes 
Transrhenanus. and our further investigations fullly confirmed our as- 
sumption, as the reader will realize by regarding our commentary upon 
the topographic details. 

The marked individuality of Prot. Ab also appears from the statistical 
classification, undertaken by Ptolemy on Germanic ground. 

There are two Germanic districts, in which the authentic towns con- 
tain numerous instances of the second class. The one is the mercantile 
road from the Danube to the Prussian Amber coast, cf. § 21, e. ; and 
the other is the region of the Limes. 

In the following, we have made a synopsis of the classification, 
according to four of the best MSS., cf. p. 69. 

Our synopsis is set forth with all reservation, as it is not always 
easy to make out the true significance of the vignettes in the various 
MSS. But at any rate, it seems to show that the distinction between 
the second and third class reflects an actual difference of importance. 
All of the 8 second class towns are situated inside the Limes, and most 
of them possess remnants of Roman fortifications. Mattiakon and Bomoi 
Plavioi are besides emphasized in various ways. The towns of the third 
class, on the contrary, are to a great extent situated outside the Limes, 
viz. Munition, Stereontion, Kanduon, and Grauionarion ; and none of them 
seems to possess noticeable remnants of fortifications. 

The Athos Atlas differs from the scheme of the other MSS., in so 
far as it emphasizes only three of the towns concerned: Amisia I cl., 
and Mattiakon and Bomoi Flavioi, II cl. This scheme is too isolated as 
to be regarded as Ptolemaic, but at any rate it reflects the geographical 
horizon of classical times. P'or the superior rank, attributed to Mattiakon 
and Bomoi Flavioi, corresponds to their actual importance, and this fact 
could scarcely have been known by a mediaeval copyist. 






Ancient supplementary 

Modern continuation; 

remnants of Roman 












Astronomic observations (Pto- 
lemy). Perhaps =: the 
fortress built by Drusus 
and restored by Germanicus 


fEms 1 
Remnants near 
[ Heftrich 






Bomoi Flavioi 



Aquae Mattiacae, fashionable 
bathing place (Ammianus) 

"Imperial Flavian altars"; the 
name translated into Greek 
(Ptolemy). Aris Flavis (Ta- 
bula Peutingeriana) 

Tenedone (Tab. Peuting.) 

(Wiesbaden) | 
.Remnants / 

I Remnants j 

Zartuna, Zarten 
fLoricha, Lorch 
\ Remnants J 

town on the river 
J Gunzenhausen ? | 
(Remnants j 


























Grinarione (Tab. Peuting.) 
Biricianis (Tab. Peuting.)? 

Groningen?, outside 
the Limes 

(Walhesdorf ?), outside 

the Limes 
Strinz, outside the 

Kohden, outside the 





















Septimiaci VII (Tab. Peuting.) 


Biburg?, outside the 


Finally, one more statistical observation must be added. As we men- 
tioned above, the towns of Prot. Ai? cluster about the mountains Abnoba 
and Albia. East of these mountains, the Ptolemaic map suddenly dis- 
closes a relatively large region without towns, only filled with displaced 
tribe-names such as Kuriones — Buroi, Chaituoroi = Chattuarioi. Then 


further east, in the region of the mountains Semanus and Sudeta, the 
towns begin again, among which the fictitious place Marobudon = 
"castellum Marobudui" (Tacitus, "Annals" II, 62), and the disguised 
mountain Bikurgion ■= Asbikurgion, Askiburgion, transplanted thither from 
north-eastern Bohemia. 

It is obvious that the empty room marks the eastern limit of Prot. Ab. 
What lies farther east, is derived from other sources, such as the Tacitean 
Annals, the Ptolemaic prototypes Bi^ B2, D etc. 

f. Occurrence of Duplicates. 
It is not possible to point out a single obvious duplicate which is 
shared by Ab. The town-names within its sphere make the impression 
af containing no fancy repetitions. Apart from solitary invaders such as 
Nouaision, they seem to betray a pure and well preserved tradition. 

g. Linguistic Marks. 

Within the relatively limited number of names, belonging to Prot. Ab^ 
we observe some noteworthy Latinisms. 

Munition, the Latin appellative "munitio" = "fortress".. 

Vulgar Latin casus obliquus -one in yiwrnWon and Grauionari<?«. Cf. 
the correspondence to the latter name on the Tabula Pentingeriana: 
Gr'msinone. Grauionarion is perhaps a misread Latin word "granary". 

Setuako-ton, = Septemiaci (VII) Tab. Peuting., seems to have been 
enlarged through misreading of the Latin figure VII. It must be noticed 
that the Athos Atlas adds the Latin figure LIIII above Tarodunon, 
whereas the Tab. Peuting. writes Tenedone XIIII. If the reading of the 
Athos MS. is no late interpolation, it must be connected with that of 
the Tab. Peuting. De^/^ona and Riusia^^^a contain the typical Ptolemaic 
transscription of Latin v, not occurring in the sections drawn from Greek 

There are no Greek marks, except the translation Bomoi Flavioi 
instead of Arae Flaviae. In this solitary case, the importance of the 
town concerned makes the translation quite natural. 

Whereas it is generally difficult to decide, whether the Latinisms 
belong to the local prototypes or to the collective one (A), the question 
in the present case seems easier to solve. Vulgar Latin forms such as 
Munition and Grauionarion are the typical mark of itineraria like the 
Tabula Peutingeriana. If the termination in Setuako-ton is to be derived 
from a Latin figure of road distance, it points decidedly towards a source 
of the same sort. 

We therefore conclude that Prot. Ab has a marked linguistic indivi- 
duality, betraying that this original map was a Latin document with the 
vulgar spelling, known from the itineraria. 


h. Literary Milieu. 

The chronological position of Prot. Ad is relatively easy to define. 
We know that the establishment of the Limes was begun under the 
Emperor Domitianus, continued under Trajanus (98 — 117 A. D.), and 
completed under Hadrianus (117 — 138 A. D.). As Al? contains at any 
rate the Vallum Trajani, it cannot have been designed earlier than about 
100 A. D. ; if it contained also the Vallum Hadriani, it would originate 
from after 117. 

Corresponding to this chronological definition, there are no traceable 
affinities with the stage of Strabo, Mela, and Pliny, who all lived before 
the establishment of the Limes. 

But the negative statement is of no great value, as the existing 
affinities with later Pre-Ptolemaic authorities are conspicuous almost ex- 
clusively by absence. 

Tacitus, it is true, mentions the establishment of the Limes, but he 
supplies scarcely any local particulars which re-appear in Prot. Ad. 

A vague affinity with Tacitus may be seen in the fact that the 
Ptolemaic place-names behind the Limes betray the predominance of 
Celtic nationality. Tacitus, *'Germania" ch. 29, states directly that the 
corner between the Rhine and the Danube had lately been occupied by 
Celts from Gaul. 

The important fortress Amisia at the north end of the Abnoba, as 
we mentioned above, may be connected with the fortress built in the 
Taunus mountains by Drusus and restored by Germanicus, see Tacitus, 
"Annals" I, 56. 

All other traceable affinities with classical authorities seem to point 
towards Post-Ptolemaic times. 

Affinity with Ammianus Marcellinus XXIV, 4: Mattiakon = Aquae 

Affinity with Vopiscus ch. XIIL Albia = Alba. 

Affinities with the Tabula Peutingeriana : 

Tarodunon (LIIII?) = Tenedone XIIIL 
Bomoi Flavioi = Aris Flavis. 

Grauionarion = Grinarione. 

Setuako-ton = Septemiaci VIL 

Riusiaua = Biricianis.?^). 

An important common element of Prot. Ad and the Tab. Peuting. 
is the knowledge of that mountain which is in the former document 
called Albia, = the present Alb. 

*) Suggested by C. Miiller, 


Also the eastern Limes, the Vallum Hadriani, may have been re- 
presented on both maps, although in disguised form. We mentioned 
above that the Ptol. constructor may have identified this part of Limes 
Map with the Sudetian mountains. On the Tab. Peuting., the Limes i& 
obviously reflected by the so-called Danube, for the towns, placed "south" 
of this river, in reality belong to the region north of it and are the 
fortresses along the Limes: Samulocenis, Grinarione, ad Lunam, Aquileja, 
Opie, Septemiaci. 

The southern part of Prot. Ab is evidently drawn from a document 
from which equally the corresponding part of the Tabula Peutingeriana 
must be derived. 

The Tab. Peuting. contains nothing corresponding to the Ptol. map 
of the middle and northern Limes. We may compare the facts that 
Ptolemy and the Tabula are most intimately related with eachother in 
northern, middle, southern, and south-western Dacia, whereas the Tabula 
lacks any sign of correspondence with Ptolemy's description of the 
eastern and north-eastern section. It seems that the selections were made 
from the source of the Ptolemaic map in an unequal manner by the 
author of the Tabula who left out entire sections for more or less 
arbitrary reasons. 

i.. Examination of Details. 

The design of the original Prototype Ab seems to have been 

Its present appearance has of course suffered deterioration through 
the Ptol. constructor, as we saw above. But even in the Ptolemaic 
disguise, several parts of Prot. Ab may still be used for the design of 
the Limes without altering a single stroke. Cf. the map accompanying 
the publications of the Limes Commission, and reprinted in Meyer's 
"Konversationslexikon", Art. 'Tfahl" *). 

We shall now try to identify the names contained in Prot. Ab, using 
as material the Ptolemaic Version I, especially Codex Urbinas 82. 

The mouth of the river Vidros, as we mentioned above, corresponds 
to the present river Wied, debouching into the Rhine at the beginning ot 
the Limes. From the linguistic point of view, the correspondence is not 
quite exact. Probably, the original form of the name was not directly 
Vidros, but at any rate ressembled this name so much that the Ptol. 
constructor was led to make a mistake. 

The so-called "river-course" of Vidros = the north-western part of 
the Limes. Notice the exact representation of the winding wall! 

^) Sarwey, Fabricius & Hettner, "Der obergermanisch-raetische Limes des Roemerreichs". 
Heidelberg, 1895 seq. 


Stereontion, town east of the "Vidros" = the present Strinz east of 
the Limes. It is subdivided into Strinz-Trinitatis & Strinz-Margaretha. 
The name appears on p. 41 — 52 of Andre's Atlas, 4th edition, where 
also most of the other modern names mentioned beneath may be found. 

River Amisias, running parallel with "Vidros" = the mediaeval Emisa, 
now Ems, running parallel with the north-western part of the Limes. 

Fortress Amisia, represented as town of the first class, with three 
towers and astronomic observations, situated directly south of the head 
of Amisias = the present Ems, situated at the head of the homonymous 
river. It corresponds to fortress no. 9 on the map of the Limes com- 
mission, at the present place called Heftrich, close to Feldberg, the 
summit of the mountain Taunus. Singularly enough, this obviously im- 
portant fortress is not mentioned directly in historical literature, but we 
may identify it with the castle built in the Taunus by Drusus and 
restored by Germanicus. Cf Tacitus, "Annals" I, 56, describing the 
undertakings of Germanicus: "posito castello super vestigio paterni prae- 
sidii in monte Tauno, expeditum exercitum in Chattos movit." — The 
map of the Limes Commission contains a fortress called Ems, registered 
as no. 4, but in reality no. 6. We suppose that this place is not the 
ancient Amisia, which ought to lie at the head of the Ems, not west of 
this river, where the fortress no. 4 (6) of the Limes is situated. 

Munition, town on the river Amisias, = a Roman "munitio", or 
"fortress". The Latin word is most likely no proper noun, but simply 
marks the place of an anonymous fortification. It may be identified 
with the mediaeval Walhesdorf, now Wallsdorf, if this place-name is to 
be translated "village of the Roman" ; but of course it is equally possible 
that Wallsdorf is founded by a German with the name Walh i. e. Roman. 
— At any rate, the existence of advanced Roman fortifications outside 
the Limes is confirmed through the excavations undertaken by the Limes 
Commission in other regions. In the neighbourhood of the "munitio" 
concerned, we also find traces of Roman population, e. g. the mediaeval 
Thabernae i. e. "taverns", now Dauborn, situated a little west of the 
river Ems. 

River Vidros, upper part = the present Wetter. The latter form is 
the exact linguistic correspondence to Vidros, according to the law of 
"High German sound-shift" ("Lautverschiebung"). 

Northern end of the mountain Abnoba = the advanced wing of the 
Limes in the Wetter district. Roman place-names like Leitcaster, now 
Leihgestern, still accompany the remnants of the Limes in these regions. 

Kanduon or Kaiduon, town east of Amisia and Abnoba, directly 
south of the western end of the mountain Melibokos, = the present 
Kohden, east of the town Ems and of the Limes (= Abnoba), and 


directly south-west of Vogelsberg which forms the western continuation 
of the mountains Ron and Thiiringer Wald (= Melibokos). 

Mattiakon, town of the second class, inside the line Vidros-Abnoba 
= Aquae Mattiacae, "the Baths of the tribe Mattiaci", now Wiesbaden, 
inside the Limes; fortress no. 31 on the map of the Limes Commission. 
The Aquae Mattiacae are mentioned by Ammianus Marcellinus XXIX, 4. 
Like the present Wiesbaden, Mattiakon seems to have been a fashionable 
place for mineral baths. 

Melokabos, town on the eastern side of the Abnoba, directly in the 
middle of its extension from the north towards the south =: the present 
Miltenberg, directly in the middle of that part of the Limes which runs 
from the north towards the south; fortress no. 38 on the map of the 
Limes Commission. Stephanus of Byzance mentions the "ethnicon Me- 
lokabenos" which would imply that the place had a certain importance, 
but according to Alfr. Holder, "Altceltischer Sprachschatz", art. Melokabos, 
the statement of Stephanus is not true. The position of Miltenberg 
corresponds to that of Melokabos, as it commands the place where the 
Limes leaves the river Main. From the linguistic point of view, the 
correspondence is not quite exact, but the geographical coincidence is 
so striking that it leaves scarcely any doubt of the identity. If Melokabos 
is a Ptolemaic metathesis of Melibokos, as R. Much suggests (see Holder, 
1. c), the change into Miltenberg would be less difficult to understand. 
The metathesis might also be of popular origin, for the German forms 
of ancient names along the Rhine, Neckar and Danube contain several 
cases of such irregularities^). 

Eastern outline of the Abnoba, between Melokabos and Lokoriton = 
the part of the Limes called Vallum Trajani, between Miltenberg and 

Western outline of the Abnoba south of Melokabos = i) the fortifica- 
tion wall between the rivers Main and Neckar, called the Miimling Line; 
2) the middle part of the river Neckar. Notice the south-eastward 
turning of the southern Abnoba, corresponding to the curving of the 
Neckar ! 

Lokoriton, town on the eastern side of the Abnoba, at the southern 
end of this mountain = the mediaeval Loricha, now Lorch, at the southern 
end of the Vallum Trajani; fortress no. 63 on the map of the Limes 
Commission. Loko-riton is a Celtic name, meaning the "Ford of Lokos", 
consequently, a rivulet or brook running through the place must have 
had the name Lokos. From the linguistic point of view, the corre- 
spondence with Loricha is not quite exact, but the geographical coin- 

^) Borbetomagus, *Borvetomagus = Wormaza, Worms; Armissa = Rems; Brocomagus 
= Brumagad, Brumpt; Alkimoenis =^ Altmtihl; Fergunnia :=r Franken Hohe. 

§ 2 1. LOCAL PROTOTYPE /I* 75 

cidence seems to exclude doubts of the identity. Cf. the linguistic 
irregularities mentioned above. 

Northern outline of the mountain Albia (perhaps continued by the 
Sudetian mountains) = the part of the Limes called Vallum Hadriani. 
Albia is the mountain called die rauhe Alb or die schwabische Alb. 

Grauionarion, town north of the Albia and east of the Abnoba = 
Grinarione on the Tabula Peutingeriana. It may be the present Groningen 
situated north of the Vallum Hadriani and east of the Vallum Trajani. 

Setuako-ton, town south-east of Lokoriton, = Septemiaci on the Tab. 
Peutingeriana. The termination -ton seems to reflect the road distance 
(VII), added after Septemiaci. 

The eastern and southern outline of the Albia would coincide with 
the north-western frontier of the Roman province called Ra^tia; Ab may 
have contained the demarkation line. The name of the Roman province 
persists till our days exactly in these regions, as Riesz, in mediaeval 
times Retia, Rezi. 

The western outline of the Albia would coincide with the upper course 
of the river Neckar. 

Southern end of the Albia = southern end of the Schwarzwald, the 
so-called Belchen, which is connected with the southern parts of the Alb. 

The Helvetian desert in Prot. Ad may have represented the same 
mountain which appeared as Abnoba on the collective orographic map of 
Germany, Prot. A, and thus would mean the Schwarzwald. In the 
Burney MS., the map represents the Helvetian desert by a long line 
running in the direction SW-NE exactly where the Schwarzwald ought 
to be situated, cf. Fig. 8. But it may not yet be regarded as certain 
that this design is of classical origin. 

Bomoi Flavioi, town of the second class, on the western outline of 
the Albia = Aris Flavis on the Tabula Peutingeriana, the present Rott- 
weil, situated on the upper course of the Neckar. As its name shows, 
the town contained a temple with altars of the Imperial Flavian family, 
and consequently must be regarded as a district capital. This degree of 
importance is reflected by the Ptolemaic vignette. It is also noteworthy 
that the Latin name has been translated into Greek. 

Tarodunon, town of the second class, north-west of Bomoi Flavioi = 
Tenedone on the Tabula Peutingeriana = the mediaeval Zartuna, now 
Zarten, south-west of Rottweil. The town occupies a central position in 
the inner valley of the river Dreisam, whereas the entrance of this valley 
is dominated by the large city of Freiburg. We may suppose that the 
importance of the classical Tarodunon was due to the same factors which 
have made the present Freiburg grow large. 

C. Miiller in his edition of Ptolemy sets forth a series of suggestions 
in order to identify the Ptol, towns within the eastern area of Prot. Ab, 

'j^ Ptolemy's maps of northern Europe 

viz. Devona — Dewangen; Kantioibis = Gunzenhausen, fortress no. 71 
on the map of the Limes Commission; Bibakon = Biburg^); Brodentia 
= the mediaeval Brenza or Prenza, now Brenz, situated on a homonymous 
river which debouches into the Danube; Riusiaua = Biricianis on the 
Tabula Peutingeriana. 

Some of the towns concerned are situated within the Albia, whereas 
the others form a fringe closely connected with this mountain. Conse- 
quently, we may take it for granted that they belonged to the special 
map of the Limes region. If Miiller's suggestions be correct, the ar- 
rangement would however have been more or less confused. As Prof. C. 
Mehlis is in near future publishing a detailed study of the Ptolemaic 
towns ("Petermanns Mitteilungen"), we think it adviseable to refrain from 
positive statements till this special research of the topographical expert 
has appeared. 

Alkimoenis on an anonymous affluent of the Danube is obviously named 
after the river Alcmona, now Altmiihl, debouching into the Danube west 
of Regensburg. But it is not absolutely certain that the town and river 
actually belonged to Prot. Ab. 

j. Conclusion. 

Prot. Ab must be called well verified both from topographic, statistical 
and linguistic points of view, partially also from the literary. Its indivi- 
duality is still more self-evident than that of Prot. Aa. 

Regarding the Limes district, Ab has the same unrivalled importance 
as Prot. Aa regarding the geography of ancient Denmark, The light 
shed by this document on the working of the Roman military topo- 
graphers must be called literally astonishing. Ab is equally important 
from the linguistic point of view, because it gives valuable information 
concerning the distribution of nationalities. We notice that the names 
inside the Roman Limes district are nearly all Celtic, the Imperial colony 
Arae Flaviae forming the only exception. Traces of German nationality 
appear on the frontier, viz. in the termination -is, added to the Non- 
German names Alkimoen(is) and Kantioib(is). Advanced Roman positions 
on German ground are marked by the names Amisia, Munition and 
Grauionarion, both of the latter showing the type of the vulgar Latin 
tongue. Cf. Fig.s 30 & 31. 

^) The name Biburg occurs repeatedly in the Danubian region. One is situated north 
of the Vallum Hadriani, a little east of Gunzenhausen; another on the southern side of the 
Danube, near the end of the Vallum. 

§ 22. LOCAL PROTOTYPES Ac, Ad. & At 'J J 


a. Summary of Contents. 

Ac is a physical map of Dacia, with probably few or no towns. 
Executed perhaps before the Roman conquest. Correctly amalgamated 
with A. Cf. Fig. 13. 

Ad & Ae are itineraries, describing Dacia; c6ntaining rivers, tribes, roads 
and towns. Ad and Ae are partially duplicates of eachother; scattered 
duplicates besides occur in Bi, B2 & F. Latin marks. Executed after 
the Roman conquest of Dacia 105 A. D. — Affinities with the Tabula 
Peutingeriana (— the Anonymus Ravennas). The prototypes seem to 
have been amalgamated before the times of Ptolemy; the map resulting 
is roughly speaking correctly amalgamated with ^. Cf. Fig.s i & 12 — 18. 

b. Ptolemaic Localisaton. 

The correct localisation of Prot. Ac was a natural consequence of its 
distinct natural outlines. The region between the Carpathian mountains, 
the lower Danube, and the Pruth, is formed by nature in such a manner 
that it lends itself quite readily as a subject of separate description^). 
For similar reasons, it was easy to incorporate the physical map Ac 
correctly with the Pre-Ptolemaic collective map of Europe. The Danube, 
as the southern and western frontier of the region mentioned, was com- 
pletely known beforehand, because it formed the frontier of the Roman 
Empire since the beginning of our era. And the large angle formed by 
this river within the region of modern Hungary offered a firm basis for 
the localisation. 

Prot. Ad and Ae are placed within Dacian territory. They are so 
far localised correctly, and in our first article on the subject^) we con- 
sequently assumed that Ptolemy's physical design of Dacia belonged to 
one of them. Through further investigations, however, we observed that 
neither Ad nor Ae agrees sufficiently with the physical map so as to be 
assigned to its original contents. This was the reason that obliged us to 
assume the existence of a separate physical map Ac^ different from Ad 
and Ae. — Prot. Ad is limited to a narrow fringe, attached to the Ptole- 
maic rivers Danubios, Tibiskos, and Hierasos. Its interior elements have 

^) The same law of geographical limitation is traceable in the extension of the Roman 
s dominion over Dacia, and much later re-appears in the establishment of the Daco-Roman 

2) "The Scott. Geogr. Mag." XXX, p. 66. 

yS Ptolemy's maps of northern Europe 

suffered displacement both towards the east and towards the west. Ae 
fills out the interior part of the Ptolemaic Dacia, evacuated by the details 
of Ad. An inexperienced editor seems to have misunderstood the southern 
outlines oi Ae, i. e. the rivers Danube and Aluta (and Theiss?), identifying 
them with the Transylvanian Alps of the physical map Ac. Thus he 
transplanted Saldensioi and Zusidava Ae [— Sallis & Sukidava Ad) from 
the southern side of the Danube to the northern, and the baths of Her- 
cules (Hydata) Ae from the Iron Gate to the interior Dacia, and so on. 
— The incorrect combination of Prot. Ad and Ae re-appears on the 
Tabula Peutingeriana which is again reflected by the descriptive text of 
the Anonymus Ravennas that often supplies a better or more complete 
reading. Cf Fig.s 15 and 16. In our research, we understand the Ta- 
bula as including the evidence of the Anonymus Ravennas, if no diver- 
gence is expressly stated. 

Even if the Ptolemaic amalgamation of Ad and Ae with the physical 
framework of Ac is incorrect, the errors generally do not assume . larger 
dimensions There are no displacements of entire provinces, and the 
parallelism of the duplicate series is in most cases undisturbed. Only a 
few names have been transplanted far away from their proper places. 
Paloda or Polonda Ad has emigrated from west of the Aluta to the 
border of the Prut (Fig. i). Sangidava Ae appears in Ad with the tri- 
plicate forms Singidava and Zargidava, the one in western Dacia, the 
other near the Dacian coast of the Black Sea. Three Danubian towns 
east of Potulatensioi Ae seem to have been moved too far east and 
placed in reverse order, viz. i Sornon, 2 Tiason, 3 Netindava, corre- 
sponding to the present 3 Soareni, 2 Teascul, i Nedeia. Cf. Fig.s 17 
& 18. We suppose that they belong to Prot. Ad^ but it cannot be made 
out exactly because they are ignored by the Tabula Peutingeriana. 

c. Definition of Limits. 

Ac may claim the entire physical details appearing on the Ptolemaic 
map of Dacia — mountains and rivers. The southern continuations of 
the Carpathian chain, lacking on the Ptol. map, seem to be traceable in 
the presumable outlines of Prot. Ae, cf. Fig. 13. 

Ad and Ae seem to have supplied almost the entire tribes and towns 
of Dacia. The two prototypes at least claim so many of these details 
that very little is left which might bo suggested as possibly belonging 
to Ac. 

The mutual relations of Ad and Ae appear from the duplicate series 
compared with the Tabula Peutingeriana. The system of routes deduced 
therefrom is summarised below, cf. the detailed synopsis under i. 

§ 22. LOCAL PROTOTYPES Ac, Ad & At 79 

Ad Ae 

Aizizis — Tibiskon (= Tabula) 

Sallis — Zurobara = Saldensioi — Ziridava; (with conti- 

nuation Ziridava — Karrodunon = 

Dierna — Zarmizegethusa (= Tabula) 

.... Pinon — Sukidava = Drubetis — Pimm — Zusidava (= Ta- 


a) .... Predav(a) — Singidava 1 

b) Sukidava — Zargidava — Petroda- > = Zusidava — Buridav(a)—Sangidava— 

va — Karsidava J Patridava (Karrodunon) (= Tabula). 

Dififerences from the Tabula may occur, but are of little import. They 
will be dealt with under the heading ^'general topographic scheme". 
Cf. also under "examination of details". 

Apart from the displacements mentioned, we observe no confusion be- 
tween Ad and Ae worth speaking of. On the Ptolemaic map, the two 
prototypes lie neatly beside eachother. Only in the south-western corner, 
they wedge themselves a little into eachother's areas, Frateria and Ar- 
V\ww2i Ad invading ^^^), and Saldensioi and Drubetis ^^ projecting corre- 
spondingly into the territory of Ad. 

The greater part of Ae seems to have been bounded by river-courses, 
viz. I. the Danube from Gran to Semlin (or eventually the Theiss), 2. the 
Danube from Semlin to Nicopoli, 3. the Aluta. Only Saldensioi and 
Zusidava, = Sallis & Sukidava Ad, and perhaps Albokensioi, cf. under f, 
belong to the southern side of the Danube. We have not been able to 
discover sure traces of Ae west or south of the above line of demar- 
kation; the further list of Cisdanubian duplicate names, collected in "The 
Scott. Geogr. Mag." XXX, p. 66, seems to be drawn from other sources, 
cf. under "duplicates". If Prot. Ae was thus limited by a line Danube 
— Aluta or Theiss — Danube — Aluta, we may assume that the duplicate 
map Ad had the same line of demarkation. 

So much for the mutual relations of Prot. Ad and Ae. As soon as 
we leave Ptolemy's map of Roman Dacia, we miss almost completely 
their distinctive marks, i. e. the duplicate series, and also the Tabula 
Peutingeriana deserts us. We therefore see here provisionally no means 
of distinguishing the exact origin of the Ptolemaic elements. We may 
only point out collectively the extreme northern outposts of Ptolemy's 
Dacian prototypes which we shall here designate as Acde for want of 
interior distinction. Cf. Fig.s 14—15. 

Outposts of Acde in these regions are the tribes Karpianoi, Tagroi, 
Biessoi, Sabokoi, Burgiones, Anartofraktoi, Koistobokoi *transmontanoi. 

^) Frateria and Arkinna = the present Fratesti and Arcan, see Fig.s 17 — 18. 

8o Ptolemy's maps of northern Europe 

The Karpianoi contrast with their alter- ego Harpioi and with the town 
Harpis. The "ethno-topic" combination of Harpioi and Harpis seems to 
indicate that these names belong to Prot. F. 

The Burgiones contrast with their alter-ego Buroi in Germania, be- 
longing to Prot. Bi, whereas the triplicate Kuriones B2 is pushed far 
away into interior Germania. 

The frontier between Acde and Bi coincides with that of the Ptole- 
maic sections Sarmatia and Germania. 

A whole series of displaced tribes from Prot. E collide with the 
north-western outposts of Acde^ viz. Ombrones, *Ouarinoi, *Burgundiones, 
*Gutones, Finnoi, cf. § 26. We may also attribute to E some invaders 
in Roman Dacia. The Ratakensioi, as C. Miiller suggests, p. 144, seem 
to be the Rakatriai Bi = Rakatai B2. Kotensioi (or Kontekoi Athos 
Atlas) = *Kotnoi, *Koteinoi of Bi & B2 (= Kytnoi in Pannonia?). The 
Teuriskoi seem to be the well-known Celtic tribe of Tauriskoi in the 
"Hohe Tauern" ; Strabo also calls them Teuristai, VIII, p. 293. 

d. General Topographic Scheme. 

The physical map Ac seems to have been of superior quality, like 
that of Germany. Cf. Fig. 13. It appears from the visible design of 
the rivers Tibiskos, Alutas, and Hierasos, but still more perhaps from 
the indirectly observed traces of the mountain system. We presume to 
have discovered them by pointing out the extension of our assumable 
Prot. Ae, for its outlines correspond too strikingly with the Transylvanian 
Alps; accidental coincidence seems to be scarcely conceivable. The like- 
ness is especially conspicuous in the south western corner where the 
Saldensioi Ae coincide with the isolated mountain chain projecting to the 
Iron Gate. The only natural explanation of this coincidence is the 
assumption that Prot. Ac contained a design of the Transylvanian Alps 
and that the outlines of Prot. Ae were identified herewith, owing to a 
misunderstanding on the part of that cartographer who amalgamated the 
two maps. 

It may be regarded as questionable whether the almost complete 
separation of the ea.stern and western Ad (cf. Fig. 14 and p. 78) is 
original or whether it is due to the cartographer who amalgamated them. 
If we are right in identifying Singidava Ad with Zargidava Ad (= Sangidava 
Ae), there would be some reason for regarding the separation of the 
sections concerned as original: Singidava and Zargidava would mark the 
same route, drawn from dififerent sources, and the author of Ad would 
have ignored the identity of both names, because he reached the station 
from two opposite points of departure. 

The system of routes, as we may reconstruct it by comparing Prot. 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae 8 1 

Ad and Ae, sometimes differs from that of the Tabula Peutingeriana. 
Here the question arises which representation is to be preferred. 

According to the Ptolemaic map, the stations Tiriskon and Argidava 
could without any difficulty be combined with the route leading from 
Hydata to Porolisson, belonging to Prot. Ae, and corresponding to the 
route Ad Aquas — Porolisso on the Tabula. But the Tabula combines 
Tivisco (= Tiriskon) with the route Tierna — Sarmategte, and Acidava 
(= Argidava) with the route Drubetis — Rusidava. The route Tierna — 
Sarmategte corresponds to a line Dierna — Zarmizegethusa in Prot. Ad, 
and this prototype would consequently claim the Ptolemaic station Tiriskon, 
if the evidence of the Tabula is to be regarded as decisive. In return, 
the Ptolemaic duplicate Tibiskon with its surroundings must then be 
assigned to Prot. Ae. 

To begin with, we actually adopted this view, owing to the fact that 
the Tabula preserves the road lines which are eliminated on the Ptole- 
maic map. But later we realized that such an arrangement is* impossible 
from the Ptolemaic point of view. We notice the following parallel of 
Ptolemaic road stations: 

Ad: Sallis — Tibiskon — Zarmizegethusa— Zurobara — Singidava. 
Ae : Saldensioi — Tiriskon — Zermizirga — Ziridava— Sangidava. 

The correspondence leaves no doubt that we are here faced with an 
original route which has been eliminated by the author of the Tabula. 
He erroneously transplanted Tiriskon Ae to a fragmentary route of Ad, 
leading from the Iron Gate to Zarmizegethusa, and he transplanted Ar- 
gidava Ae to another route of the same prototype Ad, viz. Drubetis — 

Ptolemy places Karrodunon north of Porolisson, whereas the Tabula 
has a station called Cersie south of the latter town, and south of the 
Carpathian mountains. We identify Cersie — Karrodunon with the present 
Krosno north of the mountains (cf. p. 85). Consequently, the Ptole- 
maic representation seems to be more correct. 

Dacia east of the line Pretorio — Apula — Porolisso is left blank by the 
Tabula, cf. Fig. 16. It seems, however, that the regions have not been 
completely eliminated, . but appear with wrong localisation, transplanted 
to the south-eastern side of the Danube. Next to Sucidava in Moesia, 
the Tabula places a town Sagadava, = Sancidapa Anon. Rav., which is 
ignored by Ptolemy, by the Itinerarium Antonini, and by all other 
authorities. It seems to be the Ptol. Zargidava Ad from the northern 
side of the lower Danube = Sangidava Ae. In order to explain how it 
could be transplanted south of the river, we may suggest that the author 
of the Tabula identified the neighbouring Ptolemaic town Karsidava with 
Capidava in Moesia which is known from the Itin. Antonin. and also 


82 Ptolemy's maps of northern Europe 

from an inscription (Capidavensis). If the reader compares our figures 15 
and 16, it will strike him that the Ptolemaic towns Porolisson — Napuka — 
Zargidava — Karsidava and the correspondences Porolisso — Napoca— Saga- 
dava — Calidava on the Tabula occupy fairly corresponding positions. In 
both documents, a square figure is formed. A river separates Porolisson 
and Napuka from Zargidava, and equally Porolisso & Napoca from Saga- 
dava. — Apart from Zargidava and Karsidava, no other towns from 
Ptolemy's eastern Dacia are traceable on the Tabula. The Anonymus 
Ravennas possesses an additional route running, as it seems, from the 
mouth of the river Tyras to Porolisson: Phira (Thira in Guido's Geo- 
graphy), Tirepsum, Iscina, Capora, Alincum, Ermerium, Urgum, Sturum, 
Congri, PoroUisum, Gertie. But apart from Thira, PoroUisum, and Gertie, 
the names have no likeness with Ptolemaic ones. 

It must be added that the author of the Tabula has transplanted 
about a dozen towns from the northern side of the upper Danube to the 
southern. 'Moreover, he is guilty of a really Procrustean treatment of 
an entire region about the lower Danube. The surroundings of the river 
are represented as follows: 


Hostia fl. Danubii 


The above words written with capital letters are to be read thus: LOGI 
VI REGI(S) DAG(I) PETOPORIANI, i. e. "the six places of the Dacian 
king Petoporus", and refer to a historical king Pieporus who was obliged 
to take shelter on Roman territory towards the end of the second 
century A. D. But the author of the Tabula has regarded these words 
as two tribal names, placing one half south of the Danube, and the 
other north-east of the Garpathian mountains. If he could commit such 
blunders, it is not too much to assume that he has transplanted the 
Dacian towns *Sangidava and Garsidava to Moesia. 

The result of the above considerations is that the Tabula Peuting- 
eriana shows, on certain points, a deterioration of the system of roads 
as represented by the pre-Ptolemaic map of Dacia. Here, Ptolemy proves 
superior, although his map contains no lines of roads. 

Taking it as a whole, the combined evidence of the Ptolemaic proto- 
types Ad and Ae, verified by the Tabula, speaks so distinctly that it 
enables us to reconstruct the pre-Ptolemaic system of road lines with 
approximate certainty. 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae 83 

e. Statistical Features. 

The physical map Ac contained mountains and rivers, perhaps also 
some tribes. 

Ad and Ae were itineraries. We have mentioned above that the 
western limit of Ae seems to have been the river Theiss or Danube, cf. 
p. 79. The occurrence of tribes in both prototypes would appear from 
the duplicate Biefoi Ad = Piefigoi Ae. In other cases_, the so-called 
tribes were in reality inhabitants of towns; cf. Predavensioi Ad = Buri- 
davensioi Ae = Burridava Tab. Peut. ; Saldensioi Ae = Sallis Ad, Saldis 
Tab. Peut. ; Potulatensioi Ae = Paloda Ad = Potula Anon. Rav. The 
name Albokensioi (*Albonensioi.^) evidently belongs to this class which 
besides re-appears in Moesia: Piarensioi = inhabitants of Appiaria; cf. 
C. Mtiller I, p. 444 & 463. 

We notice the absence of the "ethno-topic denomination" which 
characterizes the neighbouring prototype F. And still, there would 
have been sufficient opportunity of introducing it, as so many alleged 
tribal names are in reality simple derivations of place-names. Due 
north-east of Dacia, several instances of the "ethno-topic" nomenclature 
appear: Harpioi with town Harpis, Tyragetai along the river Tyras, 
Amadokoi with Amadokian mountains & lake and town Amadoka, etc. 

The principal contents of Ad and Ae were series of towns, connected 
by road-lines. 

In the independent northern periphery of Dacia, assigned to the 
Ptolemaic "Sarmatia", no towns are recorded. This absence of towns 
forms a contrast from the scheme of Bi which continues the town series 
towards the mouth of the Vistula on the Germanic side of the river. 

The Ptolemaic map of Dacia contains two towns of the first class, 
viz. Zarmize-gethusa and Salinai. Both are used as points of astronomic 
observation and on the map decorated with three towers ; Zarmize-gethusa 
is besides distinguished by the adjective "royal". The duplicate Zermi- 
zirga Prot. Ae misses the distinctive mark. The same representation of 
the duplicates appears on the Tabula Peutingeriana: Sarmategte with 
vignette, and Germizera without. Zarmize-gethusa is the well-known 
residence of the Dacian king Dekabalos; hence the adjective "royal". 
It may be regarded as probable that the place had some sort of distin- 
guishing vignette already in Prot. Ad. — Salinai must have been an im- 
portant saltern. It belongs to the very limited class of civil Roman 
establishments, appearing on the Ptolemaic map. The class has only two 
other representatives, viz. Hydata and Pirum (dupl. Pinon); and Salinai 
is the only establishment of industrial character. 

The Athos Atlas, differing from the context and from the Urbinas 
Atlas, assigns Salinai to the second class only, expressed by a vignette 


84 Ptolemy's maps of northern Europe 

with five battlements. This scheme is certainly not original, but still the 
variety may be derived from classical sources. At least, it is worth 
noticing that the Athos Atlas, differing from the context, places also 
Praitoria Augusta in the second class. As this town, according to its 
name, must have been an important Roman garrison, the mark of the 
second class indicates a correct knowledge of its rank. — The Tabula 
Peutingeriana represents Salinis without vignette and so far is rather 
akin to the Athos Atlas than to the ordinary Ptolemaic scheme. 

Hydata, i. e. "Baths", lacks distinctive marks in context and atlas, but 
the Ptol. description still attributes to the place a certain importance, 
appearing from the fact that its Latin name has been translated into 
Greek. On the Tabula, the corresponding town Ad Aquas has the usual 
vignette denoting bathing establishments. 

Ptolemy has another Greek translation on Dacian ground, viz. Zeugma, 
i. e. "Bridge" = Pons Trajani. It is the important military bridge 
built by the Emperor Trajanus near the Iron Gate. 

It is perhaps possible that Salinai, Ad Aquas, and Pons Trajani, 
had some distinguishing marks at the pre-Ptolemaic stage, but we must 
leave the question undecided. 

f. Occurrence of Duplicates. 

It will scarcely be necessary to point out the identity of all the 
names, indicated as duplicates of eachother on Fig. 14. In most cases, 
the identity will appear obvious from the corresponding order of the 
entire series, originally taken from itineraries. Only in some few cases, 
our assumptions require more detailed commentaries. 

The royal Dacian capital Zarmizegethusa is generally assumed to be 
different from the neighbouring Zermizirga, or Germizirga, — as the 
name is written in Codd, Paris 1403 & Vatican. Palatin. 314. C. Miiller 
re-discovers Germizirga in the town Germisara, mentioned by a Latin 
inscription ("no. 1395"), and again identified with Germizera of the 
Tabula Peutingeriana, = Germigera of the Anonymus Ravennas. The 
distance from Zarmizegethusa to Germigera seems indeed insuperable. 
But, as a matter of fact, the chasm between these apparently irreconcil- 
able forms is filled out by a large number of orthographic varieties. We 
have already mentioned the varieties of Zermizirga; those of Zarmize- 
gethusa are still more numerous. Ptolemaic MSS. : Zarmigethusa, 
Sarmisegethusa, etc.; inscriptions Zarmizegetusa & Sarmizegetusa; Tab. 
Peut. Sarmategte; Anon. Rav. Sarmazege; Dio Cassius LXVIII, 9: Zer- 
mizegethusa; cf. the river ibd. ch. 14: Sargetias, i. e. *Sar(mati)-getias. 
It is obvious that there existed several pronounciations, viz. one Sarma- 
tian, another Dacian, a third Roman, and the result was a chameleonlike 
spelling. When one name was thus spelt Zermizegethusa, Sarmazege, 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae S$ 

Sarmategte, (Sargetia-), there is practically little divergence from the form 
Zermizirga. On the Ptol. map, the two are placed close to eachother, 
and modern cartographers still reduce the distance, assuming a localisation 
which would make the one a suburb of the other from the point of view 
of a Londoner. After all we must take it for granted that these would- 
be-separate towns with almost identical names of a solitary type are in 
reality one and the same. 

Another equation which may at the first sight seem questionable is 
*Potula Ac = Paloda or Polonda Ad. The two Ptol. towns do not 
occupy corresponding positions within the duplicate series of Ac and Ad. 
And both of the differing forms seem to be confirmed by the literary 
test material: Potula is mentioned by the Anon. Ravennas, whereas the 
Tab. Peuting. contains the form Pelendoua recalling the Ptolemaic Polonda, 
However, a more detailed examination leaves no doubt that Potula of 
the Anon. Ravennas is precisely the Pelendoua of the Tabula. These 
two authorities generally register the same series of names, but there is 
a difference of arrangement in so far, as the Anon. Ravennas introduces 
a distinction between two districts, ^^Mysia", and ^'Dacia". Thus, e. g., 
the author makes a break in the route Sarmazege— *Tierna (Tema) at 
Augmonia which is the last station within the so-called district of Mysia. 
The "Dacian" part of the route is read from the opposite end, and 
when Tibis (Tibiscum) is reached the author states expressly that it is 
connected with Agmonia in the district of Mysia: "quae coniungitur cum 
civitate Agmonia patriae Mysiae". When describing the other routes, 
he does not point out the continuation from Mysia to Dacia, but in 
spite of the interrupted enumeration, no single fragment of any route is 
omitted. We are thus able to state that the Peutingerian series Romula, 
Castris novis, Pelendoua . . Drubetis is rendered by the Anon. Ravennas 
thus: Romula, Canonia, Potula, Bacaucis. Canonia is evidently a mis- 
understood abbreviation Ca. noua = Castra nova, and the following Po- 
tula must be identical with Pelendoua, at the same time coinciding with 
the place of the Potulatensioi on the Ptol. map. 

Our equation Karrodunon Ac = Karsidava Ad is supported by the 
Tabula which replaces Karrodunon by Cersie = Gertie of the Anon. Rav. 
It seems to be the present Krosno north of the Carpathian mountains, 
cf. p. 8i. 

Singidava Ad, Zargidava Ad, and Sangidava Ae, seem to represent 
a case of triplication. Zargidava = Sagadava of the Tabula, Sancidapa 
of the Anon. Rav.; Sangidava =: Acidava of the Tabula, and Sacidava of 
the Anon. Rav. We have discussed on p. 80, how it can be explained 
that the author of Prot. Ad repeats the name of the station. The dis- 
placed localisation of Sagadava on the Tabula is pointed out on p. 81. 

Argidava Ae has no Ptolemaic duplicate. When C. Miiller places the 

86 Ptolemy's maps of northern Europe 

name near the Theiss, i. e. within the area of Prot. Ad, it is due to a 
conjecture of Wilberg's. But on the Tabula, the place actually belongs 
to the duplicate series, appearing as Arcidava near the Theiss = Prot. 
Ad, and as Acidava near the Aluta = Prot. Ae. The form Acidava is 
different from its above-mentioned namesake which is a mutilated form 
of Sancidava == Sangidava Ptol. 

In "The Scott. Geogr. Mag." XXX, p. 66, we assumed that the 
duplicate series of Ad & Ae continued west of the middle Danube, 
finally reaching the northern corner of the Adriatic. They would con- 
tain, e. g., two Mursella in Pannonia, Sirota = Sisopa ibd., and in Istria 
Alvona = Alvon, i. e. the ancient and present Albona. These cases 
are perhaps too scattered to form a solid basis for assuming the con- 
tinuation of the two prototypes west of the Danube. But the duplicate 
Sallis Ad = Saldensioi Ae at any rate shows that they contained some 
parts of the Cisdanubian provinces, viz. the Pannonian district round the 
inferior Save*). 

The possibility is not excluded that the so-called Dacian tribe Al- 
bokensioi north-west of Saldensioi may be a misreading of *Albonensioi. 
In this case, it would belong to Ae and its duplicate would be Alvona, 
belonging to Ad, whereas Alvon would be a triplicate form, derived from 
another prototype. Neither the Tab. Peuting., nor the Anon. Ravennas, 
it is true, connect Saldis and Albona through a direct route. But on the 
Tabula, a route from Saldis to Aquileja almost touches Albona (Alvona), 
and the Anon. Ravennas represents Albona as the starting point of an 
lUyrian route (p. 224, ed. Pinder & Parthey). 

The duplicates which the surrounding prototypes have in common 
are so few that they do not contribute essentially to illustrate the 
making of the Ptolemaic Dacia. We have noticed: Karpianoi = Harpioi 
F, and Burgiones = Lugoi Buroi Bi, Kuriones Bi. Cf. § 26 & 23. 

g. Linguistic Marks. 

Latinisms prevail in Dacia and its surroundings. 

Dacia: the Latin word Salinai; Firum; A^^ustia, Sa;2^idava, Singi- 
dava, Zar^idava (< *Za;s5^idava). 

Sarmatia: the Latin adjective *transmontanoi ; the Latin termination 
in Karpi^«oi; Pie«^itai. 

Moesia: Karsum, Siw^idunon. 

Pannonia: the Latin dative plur. in Salk*.y; two Akumi«/^on, Akvi;^>^on. 

^) The following duplicates suggested in Pannonia inf. are questionable: Lussonion — 
Lugionon, Berbis — Serbition. Karrodunon in Vindelicia is no duplicate of its Ptol. name- 
sake in Pannonia, but must be amended into Parrodunon, cf. C. Miiller, I, p. 284. 

§ 22. LOCAL PROTOTYPES Ac, Ai & At 8/ 

There are no typical Greek marks. The translation of Ad Aquas into 
Hydata, and of Pons into Zeugma may have been undertaken by the 
Ptol. constructor. 

The presence of Latinisms and the absence of Greek marks forms a 
contrast from the sphere of Prot. F, 

h. Literary Milieu. 

The physical map Ac may originate from the first century of our 
era. The Romans would have been able to draw a "blind" map of 
Dacia before actually conquering the country : this fact appears sufficiently 
from Prot. A, i. e. Ptolemy's excellent physical map of unconquered 
Germany. Already before our era, the Romans knew the dimensions of 
Dacia, as it is stated by Agrippa in his Commentaries: "Dacia, Getica 
finiuntur ab oriente desertis Sarmatiae, ab occidente flumine Vistula, a 
septentrione Oceano, a meridie flumine Histro. quae patent in longitudine 
milia passuum CCLXXX, in latitudine qua cognitum est milia passuura 
CCCLXXXVI"; cf. Mullenhofif's "Germania antiqva", p. 49. — And 
about this time, Dacia was regarded almost as a dependency of Rome, 
see Strabo's Geography VII, p. 305, written in the first decades of 
our era. 

The itineraries Ad and Ae necessarily must represent a later stage. 
As they contain the names of Imperial garrison cities, such as Praetoria 
Augusta, it follows naturally that the date of origin should be later than 
the Roman conquest of Dacia 105 A. D. 

We have mentioned above that the combination of Prot. Ad and Ae 
re-appears on the Tabula Peutingeriana. The next question is to define 
the relations of these maps more exactly. 

There must be a prehminary statement of three alternatives: — 

1. The original itineraries might have been combined independently 
by the Ptol. constructor and the author of the Tabula. 

2. The Ptolemaic map of Dacia might be the source of the corre- 
sponding section of the Tabula. 

3. The Ptolemaic map of Dacia and the corresponding section of 
the Tabula might be derived from a common source in which the 
original local itineraries were already combined. 

Alternative no. i may be regarded as excluded. The map of Dacia 
in its Ptolemaic shape agrees too well with that of the Tabula. Even 
if the towns mentioned are not always the same, not one route of the 
Tabula is omitted on the Ptolemaic map of Dacia ^). And the Tabula has 

^) The Anon. Ravennas contains one additional route, Phira — Gertie, see p. 82. 

88 Ptolemy's maps of northern Europe 

several Ptolemaic duplicates, viz. Tivisco — Tivisco, Sarmategte — Germizera, 
Sucidava — Rusidava, Sagadava — (S)acidava. 

Alternative no. 2 is equally excluded. For the Tabula contains a 
system of road lines which does not appear on the Ptolemaic map. The 
lines, it is true, are not always drawn correctly, but the general coin- 
cidence with the Ptolemaic arrangement of towns is unmistakable and 
thus points towards inheritance from an older source. 

Alternative no. 3 is preferable. The Ptolemaic map of Dacia and the 
Tabula are co-ordinate descendants of one large original .map which 
already contained the prototype Ac, Ad and Ae in amalgamated form. 
Both of the descendants preserve certain individual features of the 
original: Ptolemy has -the relatively correct physical design and the larger 
number of duplicates, whereas the Tabula has the road -system. It must 
be added that the Tabula seems to have been influenced by the editorial 
scheme of certain Ptolemaic MSS. Germizera of the Tabula recalls the 
reading Germizirga in the Codd. Paris 1403 & Vatican. Palat. 314, instead 
of Zermizirga; Pelendoua of the Tabula reminds Polonda in the Cod. 
Vatican. 191, instead of Paloda. 

After we have so far pointed out the genetic relations of the Ptole- 
maic map and the Tabula, we may try to investigate the editorial chro- 
nology still more exactly by means of the nomenclature. 

Plrst stage. The physical map Ac, probably designed before the 
Roman conquest of Dacia, and containing no detailed nomenclature. 

Second stage. A pair of itineraries Ad & Ae, duplicates of each- 
other, describing the lately conquered regions along the Danube, the 
Theiss and the Aluta; containing one important garrison city, Praetoria 
Augusta, and one more station with a Latin name, Pirum Ac (= Pinon 
Ad); otherwise, the nomenclature is at this stage purely Dacian. 

Third stage. The originally identical series af Ad & Ae are enlarged 
with individual characteristics. Those oi Ae denote the constant spreading 
of the Roman nationality, appearing in the names SaHnai and Hydata = 
Salinis and Ad Aquas on the Tabula. Ulpianon, probably belonging to 
Ae, is the garrison city of a Cohors Ulpia. Perhaps, Ad was at this 
stage enlarged with the station *Pons (= Ptolemy's Zeugma). — The 
most important enlargement since stage II is the continuation of the 
route Saldis — Ziridava to Porolisson and through the Dukla defile to 
Karsidava (Karrodunon) north of the Carpathian chain. This is a well- 
known military and mercantile road, partially built by the Cohors Ulpia, 
as stated in an inscription. Cf. under '^examination of details", p. 94. 
It is a natural development that the individual contents of Prot. Ae advance 
most con.spicuously in the northern regions. 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae 89 

Fourth stage. The prototypes Ac, Ad, and Ae, are amalgamated. 
The road-system of Ad and Ae is still preserved. It is questionable 
whether the process of amalgamation should be attributed to Marinus or 
to a predecessor of his. 

Fifth stage. The amalgamated map Acde is incorporated with the 
Ptolemaic atlas. The road-system is eliminated. The nomenclature still 
remains chiefly Dacian. 

Sixth stage, post- Ptolemaic. The amalgamated map Acde is in- 
corporated with the prototype of the Tabula Peutingeriana. Some 25 
new names are introduced, almost all of Latin origin. The additions 
contain only some three names of Dacian origin, viz. Bersovia, *Cebonie, 
Arutela. Bersovia, a station on the present river Berzava, was already 
mentioned by the Emperor Trajanus, and so it may be a mere accident 
that the other non-Ptolemaic names of Dacian origin are not preserved 
in any documents dating from before the times of the Tabula. *Cebonie 
(Cedonie Tabula) is the present important town Cibin or Szeben on a 
homonymous river. Arutela may be a mutilation of a Latin *Ara 
Tutelae, according to C. Miiller, I, p. 447^). It is evident at any rate 
that the Dacian map of the Tabula has been completed after the final 
triumph of Roman nationality. 

It remains to discuss the provenience of the Ptolemaic tribes Koisto- 
bokoi *transmontanoi, Biessoi and Sabokoi in independent Dacia north 
of -the Carpathian mountains. The Koistobokoi fought against Rome in 
the Marcomannian war, according to Julius Capitolinus, Bell. Marcom. 
ch. XXII. The Biessoi and Sabokoi probably did the same, according 
to Miillenhoff's emendation of the corrupt names "-bessicobotes" in the 
list given by Julius Capitolinus. Thus the part concerned of the Ptole- 
maic map would seem to contain elements which were partially unknown 
to the Romans, before the Marcomannian war burst out, i. e. 166 A. D. 
Under this presumption, the elements concerned could not have belonged 
to the stage before Ptolemy, but would have been introduced by him- 
self. On the other hand, the possibility is not excluded that the said 
Dacian tribes should have become known to the Romans even earlier, 
owing to the intercourse on the mercantile road to the Prussian amber 
coast since the age of Pliny. We must leave the question unsettled. 

^) The name Brucia on the Tabula sounds non-Roman, but it is an illusion as appears 
from the correct spelling Brutia, preserved by the Anon. Ravennas. 


i. Examination of Details. 

It remains to comment upon the details of the Ptolemaic Dacia and 
Jazygia according to their positions within the system of routes. 

In order to investigate the details of the Roman routes, C. Miiller 
lays great stress on the road distances indicated by the Tabula Peutin- 
geriana. We cannot admit this valuation as quite justified, so far as 
Dacia is concerned. For the Tabula, as we have shown above, derives 
its description of Dacia from a map which already contained the proto- 
types Ad and Ae in the incorrectly amalgamated form. Moreover, the 
Tabula adds to the confusion. Names such as Rusidava, Tivisco No. 2, 
A(r)cidava No. 2, are introduced at wrong places, thus disturbing the 
road measurements concerned. Cersie and Porolisson seem to be inter- 
changed, etc. And whereas the final editor of the Tabula might easily 
correct measurements within all the then existing provinces of the Em- 
pire, he was prevented from undertaking such corrections in Dacia, be- 
cause this province had been lost to the barbarians for a full century, 
when the Tabula was published. — Under such circumstances, we regard 
it as provisionally impossible to use the Dacian figures of the Tabula as 
the basis for definite calculations. No positive results can be extracted 
from them, until the genetic relations of Ptolemy's map and of the 
Tabula have been thoroughly examined. 

After these preliminary remarks, we shall give a general synopsis of 
the routes concerned and then proceed to the examination of particulars. 
(See Tab. p. 91.) 





< g 


<^ ^ 



S^ <1 

^ < 

§ <3 

fS < 



-^ ^ ^ 

aT .2 



e 1 

< i <i 

-^ < 

U cu 


cs" «r 

<1 '3 <1 




a &- 


CJ «S 

,-~v > 

^ ^ 

. <] <1 ^ 

•^ <1 <] -^ 

i s^ 

S < < s 

.2 w 

U < <1 fe4 

1 2 

. . 

« rt 

1 t t 1 

0, Ph 

j\ ^ /t 

^ .«'~ 

/^ ^ ^ 

.2 rt 



•S .5 

r<) CO 


13 13 


c« C/i 


J- |_^ 




-^ — ' •>— ' 

d § 



1- 1 

« ^ 
3 2 



<: < 

a a* 



■ a 

< < . . 



■ % 

«r rt rt «r 








: :5 

• N 





<U I- 























d Aqua 








0^ .2 

1 'i 





eS • 






'rt' ?" 












.2), (Acidav 













n PQ cu • 1 





8 § 
.2 -^ 






• t '5 : 












a, Pie 
ra), Bie 
a, . . . 




> c3 > 1 









rt c<t 
















w C 









v» "^ ^3 





H ^ 






^ ^ ^ 

fl C .Si 

rt tuo 



S Q £ 







1— T 




u - 






I. More remoter stations of correspondence, communicating with 
the routes I, i, I, 2, I, 3 (and eventually with II). 

We mentioned on p, 86 the possibility that the present town Albona 
in Istria belonged to the prototypes Ad & Ae, as Alvona Ad and Albo- 
kensioi Ae = Alvona Tabula, Albona Anon. Ravennas. Certainly, the 
Tabula does not connect Alvona directly with the system of Dacian 
routes. But the Anon. Ravennas at least represents Albona as the 
starting point of an Illyrian route. Apart from its occurrence as Alvona 
Ad and (?) Albokensioi Ae, a third Ptolemaic prototype recorded the 
place as Alvon. Undoubtedly, Albona possessed a certain importance, 
still to-day reflected by the fact that it is the one of the two sole 
surviving Roman towns on the east coast of Istria amidst a population 
of immigrated Slavs. The gulf of Quarnero, on which Albona is 
situated, is the one main entrance to the road leading down the Save 
valley, the most direct route from Italy to Dacia. Such circumstances 
make it easily conceivable that Albona has, as it seems, become the 
starting point of the western systems of routes in Ad and Ae. 

If the occurrence of Albona in Ad and Ae is still questionable, it is 
all the more certain that a nearer starting point of the western Dacian 
systems of roads was formed by Sallis Ad = Saldensioi Ae = Saldis 
Tabula & Anon. Ravennas; i. e. *Saldae in grammatically correct Latin. 
The town was situated on the southern border of the inferior Save and, 
according to the Tabula, directly connected with that route which crossed 
the Danube, entering Dacian territory near Arcidava. Saldae is mentioned 
nowhere except by the four authorities mentioned, but its appearance in 
Prot. Ae is sufficient to prove its character as a starting point. 

2. Route I, I. 

Tab. Saldis, Arcidava, Azizis, Tivisco no. i. 
Ad Sallis, Aizizis, Tibiskon. 

After Saldis, the next main station is Viminatio, according to the 
Tabula, i. e. the well known city of Viminacium in Moesia superior, due 
east of the mouth of the Morava. 

Arcidava Tab. follows directly after the route has passed the 
Danube. This name is lacking in Ad, but its duplicate Argidava = 
Acidava Tab. appears in Ae, belonging to route I, 2, and situated at 
a considerable distance east of Tiriskon i. e. Tivisco no. 2, Tab. The 
order is reversed: Arcidava, Tivisco in I, i, Tiriskon, Argidava in I, 2; 
probably, it is the Tabula that is mistaken. 

Aizizis Ad = Azizis Tabula. The Emperor Trajanus writes that 
he went from Berzobis to Aizis, cf. Priscianus VI, p. 682 (*'Auctores 
gram. Lat.", ed. Putsch). Consequently, Aizizis must lie in the neigh- 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae 93 

bourhood of the present river Berzava, debouching into the Temes from 
the south. According to an inscription, a god named Azizus was wor- 
shipped in the Dacian town Patavissa, cf. C. Miiller, I, p. 449. 

Ti bisk on Ad, Tivisco no. i Tab.^ re-appearing as Tiriskon Ae, 
Tivisco no. 2 Tab., belonging to route I, 2. It is probably the present 
Temesvar. Ptolemy erroneously attributes the name Tibiskos to the 
river Theiss, Hungarian Tisza, whereas it is in reality preserved by the 
river Temes. 

Ptolemy places Tibiskon south of Aizizis, almost at the mouth of 
the Temes. We suppose that this localisation is due to the general 
displacement of Prot. Ad in the region concerned. The true sequence 
seems to be: i. Viminakion, south of the Danube; 2. Aizizis, near the 
Berzava, i. e. south of the Tibiskos; 3. Tibiskon = Temesvar. 

3. Route I, 2. 

Ad Sallis, Tibiskon, Zarmizegethusa, Zurobara, Singidava. 

Ae Saldensioi, Tiriskon, Argidava, Zermizirga, Ziridava, Sangidava. 

This route may be regarded as a continuation of I, i, yet with a 
partially altered line: the stations Tibiskon and Argidava re-appear, 
whereas Aizizis is omitted. 

The first station on the continued route is Sarmisegethusa (Zarmi- 
zegethusa) Ad^ Zermizirga Ae = respectively Sarmategte and Germizera 
of the Tabula. It is the terminal station of route II, capital of the 
Dacian king Dekebalos and hence called "royal" by Ptolemy; point of 
astr. observation, Ptol.; vignette with towers, Ptol. & Tabula. The name 
signifies a racial mixture of Sarmates and of Getes, i. e. Dacians; it is 
besides attributed to the river Sar(mati)-Getias, the present Sztrigi or 
Streiu, which flows past the town, cf. p. 84. 

Zurobara Ad, Ziridava Ae, next station. Perhaps the present 
Szerda hely east of the river Sztrigi. Ziridava is the right spelling. 

Singidava Ad, Sangidava Ae, terminal point of correspondence, 
otherwise belonging to route IV. Cf this route. 

4. Route I, 3. 

Tab. (Saldis), (Tivisco no. 2), (Acidava), ad Aquas, Germizera, 

Ae Saldensioi, Tiriskon, Argidava, *Aquae, Zermizirga, Ziridava, ' ^ 

Tab. Apula, Salinis, Patavissa, Napoca, Cersie, Porolisso, AAAA 

Ae Apulon, Salinai, Patruissa, Napuka, Porolisson, AAAA, Karrodunon. 

The larger part of this route is identical with I, 2. But from Apulon 

I, 3 continues due north, whereas I, 2 turns towards the north-east in 


the direction of Singidava — Sangidava. And the Tabula does not make 
the route start from Saldis or from any other western point of corre- 
spondence, occurring in Ae, but places the starting point within Dacian 
territory, viz. at the station ad Aquas = Hydata Ae. In our opinion 
the route can scarcely have contained both Tibiskon and ad Aquas; the 
beginning must be: either Tibiskon — Zarmizegethusa, i. e. from the 
Danube along the Temes to the Sztrigi; or ad Aquas — Zarmizegethusa, 
i. e. from the Danube (Iron Gate) along the Cerna to the Sztrigi. The 
exact coincidence between Ae and the Tabula makes it most plausible 
to conclude that the route leading to Porolisson started practically at 
ad Aquas, even if it were thence connected with the more remote station 
of correspondence Saldae, belonging to the routes I, i and I, 2. 

The starting point Hydata Ae — ad Aquas Tabula must be placed 
in the immediate neighbourhood of the Danube. It is the only Dacian 
bath mentioned by Ptolemy and on the Tabula; also the translation of 
Aquae into the Greek Hydata points to a certain importance. The 
place must be identical with the Aquae Herculis near the mouth of the 
river Cerna, known as a fashionable bathing establishment of antiquity. 

After passing Zarmizegethusa and Ziridava (see route I, 2), the next 
station is Apulon Ae, Apula Tabula. It is the junction with route IV 
and has a vignette with two towers on the Tabula. The town is a 
district capital after which one of the three Dacian provinces of Rome 
is called Apulensis. It is supposed to be the present Karlsburg, Karoly 
Fejervar in Hungarian. 

Salinai X^ = Salinis Tabula. A Roman saltern. Point of astronomic 
observation, Ptol. Vignette with towers, Ptol. (The Athos Atlas has a 
vignette of the second class only, with 5 battlements). According to 
C. Miiller, I, 447, Salinai was situated at Felvincz which means "saltern" 
in Hungarian. Here an inscription of the 5th Macedonian legion has 
been found. Others prefer the localisation near Thorda which also 
possesses a saltern. 

Patruissa = Patavissa Tabula & inscription; more frequently in in- 
scriptions Potaissa. According to C. Miiller, I, 446, situated at the 
present Thorda. 

Ulpianon, Garrison-city of the Cohors I Flavia Ulpia that built 
the road between Patavissa and Napuka in the year 109 A. D., according 
to a local inscription, cf. C. Miiller, I, 446. The Ptol. map places Ulpianon 
at a considerable distance west of the route. Perhaps this is an error. 

Napuka = Napoca Tabula & inscr., designated with two towers on 
the Tabula. Roman colony according to Ulpianus, "De censibus", I 
("Digesta" L, 15, i, 8), situated ten millia passuum from Patavissa, 
according to a mile-stone. The present Klausenburg, according to C. 

§ 22. LOCAL PROTOTYPES Ac, Ad Sc At 95 

Porolisson = Porolisso Tabula, Paroliss- and Paraliss- in inscrip- 
tions; — the extreme northern station in Roman Dacia south of the 
Carpathian mountains; designated with two towers on the Tabula; capital 
of one of the three Dacian provinces, Parolissensis. According to C. 
Miiller, the town was situated at the present Mojgrad where an amphi- 
theatre was built in the year 157 A. D. (inscr. n. 836). Other scholars 
are of a different opinion. We should prefer to place Porolisson farther 
north, because it is — together with *Cersie — represented as lying at 
the northern extremity of a route leading from the Black Sea up the 
river Dnjestr, according to the Anon. Ravennas (cf. Fig. 17). It might 
have been situated at the most northerly point of the river Theiss, 
which is a dominating strategical position a little south-east of the Dukla 

Karrodunon Ae, Karsidava Ad = resp. Cersie and Calidava Tabula, 
a town in the extreme northern part of route I, 3, north of the Car- 
pathian mountains, belonging to "Sarmatia", i. e. outside Roman Dacia. 
It seems to be the present Krosno that lies due north of the im- 
portant Carpathian defile of Dukla through which the route passes from 
Hungary to the upper Vistula. The form Karrodunon is Celticized, 
owing to analogy with a well-known station on the mercantile road from 
the middle Danube to the lower Vistula. Karsidava is the right spelling, 
which may also have been abbreviated into *Karsion, cf. Cersie on the 
Tabula. It is the abbreviation which survives in the present Krosno. 

5. Route II. 

Tab. Tierua, Pretorio, Agnavie, Sarmategte. 

Anon. Ravenn Tema, Pretorich, Agmoniaj Sarmazege. 

Ad Zeugma, Dierna, Frateria, Arkinna, Akmonia, Sarmisegethusa. 

The route starts and runs a little east of I, 3. We may supplement 
its particulars by the aid of the Tabula. 

The Ptol. starting point is Zeugma Ad = Pons Trajani, a military 
bridge built by the Emperor near the Kasan defile; but the Tabula 
rather starts from Dierna Ad = Tierua Tabula, Tema Anon. Ravenn., 
a well-known Roman town at the mouth of the present river Cerna. 
The Roman town Trans-Tierna seems to survive as the present 
Cerneti or Tschernetz. Tierna lies directly at the famous Iron Gate 
of the Danube, a place of high military importance^ as showh by the 
large inscription of the Emperor Trajanus ("Trajanstafel"). Thus it is 
easily understood that a cartographer should designate it as the starting 
point of a route. 

Arkinna Ad = the present Arcan, a station of route II. The Ptol. 

96 Ptolemy's maps of northern Europe 

map places it almost correctly near the river Rabon, i. e. the present 
Jiul. Only it ought to lie south of the following station, not north. 

Frateria Ad = Pretorio Tabula = the present Fratesti, a station of 
route II, placed almost correctly by Ptolemy, only south of Arkinna, in- 
stead of north. The Tabula has distorted the Dacian name into the 
Latin Pretorio, known from an important garrison city of Dacia. Although 
the place is nowadays only a village or borough, it seems to have been 
more important in past times, as the surrounding valley has been named 
after it: Val Fratestilor. It is also situated near the point where route II 
joins an important route coming from the present Rimnik on the river 

Petri s Tabula, surviving till our days as Petrilla and Petroseni, names 
of two places near the Vulcan defile, where the route leaves Roumania 
and enters Transylvania. The name is Latin, originating from the sur- 
rounding high mountains one of which is still called Petri. 

Sarmisegethusa, junction with the routes I, 2 & I, 3. The corre- 
spondence Sarmategte is represented as the terminal station on the 

6. Route III. 

Tab. DrubetiSj Amutria, Polonda (Anon. Ravenn. Potula), Rusidava. 
Ae DrubetiSj Amutrion, Potulatensioi, Zusidava. 

In the list below, we add some names from the same regions, occur- 
ring in Ad{}), which seem to have no correspondences with other 

The starting point of the route is Drubetis Ae and Tabula = 
Drobeta in the Notitia Dignitatum (5th century). It seems to have been 
a Roman fortress or bridgehead near the Iron Gate. As it appears still 
in the Notitia Dignitatum, it may have been held by the Romans even 
after they had given up the rest af Dacia. 

Amutrion Ae = Amutria Tabula. The present Motru at the point 
where the homonymous river debouches into the Ptolemaic Rabon (C. 
Miiller). The town is also called Gura Motrului. Its position at the 
river- junction gives it a certain importance. 

Netindava Ad} = the present Nedeia on a homonymous lake close 
to the Danube (C. Miiller). 

Tiason Ad} = the present Teascul on the Ptol. river Rabon, near 
the Danube. C. Miiller writes the name Tiasul, but the above ortho- 
graphy is reported to be more correct. 

Sornon Ad}. Probably the present Soareni east of Teascul, near 
the Danube. 

Potulatensioi Ae, Paloda or Polonda Ad = Pelendoua Tabula, 
Potula Anon. Ravenn. The present Potel on a homonymous lake with 

§ 22. LOCAL PROTOTYPES Ac, Ad & Ae 97 

a homonymous outlet into the Danube. Situated close to this river. C. 
Miiller identifies Polonda with the present Palitula (read: Palilula) on the 
Ptol. river Rabon, but we have seen above that the Ptol. name is simply 
a duplicate of Potula-. 

Romula Tabula, according to an inscription situated at Turnu Ma- 
gurelli facing the present district of Romaniti near the mouth of the 
Aluta. Cf. C. Miiller, I, 447. The name Romula is Latin and of later 
orgin than the Ptolemaic map which retains an almost purely Dacian 
nomenclature. We mention it here, because it marks the point where the 
detailed description of the route ceases. 

Sukidava Ad, Zusidava Ae = Sucidava and Rusidava Tabula, the 
terminal point of the route in Ae. It is the well-known town Sucidava 
in Moesia inferior, i. e. south of the Danube, at the point where the 
river suddenly turns from an eastward direction towards the north. 

The detailed description of the route really does not extend farther 
than the river Aluta, whereas Sukidava — Zusidava is only regarded as 
a far-off terminal point. The constructor of the Ptolemaic map, however, 
regarded the names from west of the Aluta as representing the entire 
space down to Sucidava and thus displaced them considerably. At the 
same time, the order of the names Netindava, Tiason, Sornon seems to 
have been disturbed, this series being turned the wrong way, east-west 
instead of west-east. Cf. p. yS (b). 

7. Route IV. 

lab. Rusidava, Burridava, Pretorio, (S)acidava, Apula, (I, 3 > Cersie A AAA ) 

Ae Zusidava, Piefigoi, Buridavensioi, Praitoria Augusta, Sangidava, Apulon, (I, 3 > AAAA Karrodunon) 

Ad (Zurobara), Biefoi, P(u)redavensioi, Singidava, ( >> AAAA) 

Ad Sukidava Angustia, Zargidava, Karsidava AAAA. 

This route seems to have been doubled in Ad, its two replicas being 
transposed respectively from the east to the west. The two fragmentary 
routes in Ad supplement eachother so as to give together the sum total 
of the route in Ae; only Sangidava Ae is twice repeated in Ad. 

Sukidava Ad = Zusidava Ae appears as the starting point of both 
routes, exactly as Zusidava appears as the terminal point of route III. 
It must, however, also here be regarded only as a far-off station of 

The real starting point, according to the Tabula, is Ponte Aluti, 
which must be placed near the Danube, not far from Romula of route III. 
It is another Latin name, later than the Ptolemaic stage. We mention 
it here in order to show that route IV starts from the mouth of the 
Aluta, not farther east. 

Piefigoi Ae, tribe south of Buridavensioi = Biefoi Ad south of 

98 Ptolemy's maps of northern Europe 

Buridavensioi Ae, Predavensioi Ad = Burridava Tabula, immedi- 
ately after Ponte Aluti. It may be the present Burdea near a homo- 
nymous affluent of the river Vede. Perhaps, the river is named after the 
town, like Ogost on the opposite side of the Danube, debouching at the 
town Augustai. Cf. p. 102. 

(Castra Trajana Tabula may be the present Troian or Traian, 
which lies however south of Burdea, not north.) 

Pi rum ^^, Pinon Ad^ belonging to the class of early Roman establish- 
ments in Dacia. Perhaps the present Pirlita, or the present Pires. Both 
of these towns or boroughs are situated north of Bucuresti. 

Komidava or Ramidava, no longer traceable. 

Praitoria Augusta Ae, Angustia Ad = Pretorio Tabula. An im- 
portant Roman garrison city, and consequently distinguished with 5 
battlements in the Cod. Athous Vatopediensis. C. Miiller, I, 447, places 
it directly on the Aluta and its affluent Govori, but if our interpretations 
of Buridava and Castra Trajana are correct, Praitoria would ratlier be 
situated a little east of the river. C. Miiller regards Angustia as the 
same place which was with a semi-Greek name called Caput Stenarum 
(Anon. Ravenn.), as both names would mean "defile". We prefer to 
identify Angustia with Augusta, as the occurrence of duplicates is so 
usual in this part of Ptolemy's Dacia. 

Cedonie Tabula, after Stenarum, must be amended into *Cebonie 
(C. Miiller). It is the present Cibin or Szeben, in German called Her- 
mannstadt, an important Transylvanian town, situated on a homonymous 

Sangidava Ae, Singidava & Zargidava Ad = Acidava & Sagadava 
Tabula (Sacidapa & Sancidapa Anon. Ravenn.). (S)acidava follows next 
Cedonie. According to Ae, Sangidava would be situated north-east of 
Praitoria. We may suggest an equation with the present Seges-var or 
Schassburg, Roum. Sighisora, situated on the river Kokel. 

Kaukoensioi, tribe a little south of Sangidava = inhabitants of the 
Caucalandensis locus which is mentioned by Ammianus Marcellinus 
XXX; 4. C. Miiller suggests that the name may be connected with the 
river Kokel or Kiikiillo which passes Seges-var in a south-westerly direc- 
tion. It has given its name to the town Kiikiillo-var or Kokelburg, the 
capital of a homonymous district. The fact that the invading Goths 
formed a district name Caucaland in their own tongue seems to attri- 
bute to the Kaukoensioi a certain importance, and hence it would be 
likely that their name might still survive. 

Patridava Ae, Petrodava Ad, no longer traceable. (Cf. addition 
p. 102). 

Markodava Ae. C. Miiller, I, 447, suggests the alteration into 
*Marodava and interprets the name as "town on the river Marisia". It 

§ 2 2. LOCAL PROTOTYPES Ac, Ad 9[ Ae 99 

might be the present Maros Ujvar. At any rate, the place must be 
localised in the region of the river Maros. 

Apulon Ad = Apula Tabula, junction of the routes IV and I, 2. 

The routes here dealt with may, in most cases, be regarded as suffici- 
ently verified, partially through the mutual correspondence of the proto- 
types Ad and Ae, partially through the supplementary evidence of the 
Tabula Peutingeriana. 

In a number of cases, however, we have commented on names of 
more questionable provenience, e. g. the town series Netindava, Tiason, 
Sornon, or the tribal name Kaukoensioi. For practical reasons we thought 
it most convenient to deal with such matters in connection with the 
routes passing the immediate neighbourhood. 

8. Details from independent Dacian and Jazygian regions. 

The Ptol. maps of south-western Sarmatia and of Jazygia contain a 
series of names which must, to a great extent, have been extracted from 
descriptions of mercantile roads. As we mentioned above, it is provi- 
sionally not possible to distinguish whether they belong to Ad or to Ae 
and their connection with the Roman system of roads is equally un- 

The tribes Biessoi, Piengitai, Sabokoi, Arsietai, Burgiones, Anarto- 
fraktoi are placed in a row from the south towards the north close to 
the frontier of Sarmatia and Germania. As the frontier-line is the dis- 
guised expression of the mercantile road from Carnuntum to the Prussian 
Amber coast (cf. § 23), the tribes concerned probably belong to a 
description of this route. All of them seem to be Dacian except the 

The Biessoi and Sabokoi are by Miillenhoff identified with the 
"sosibessicobotes" who appear among the enemies of the Romans in the 
Marcomannian war, according to Julius CapitoHnus ch. 22; read: ''Osi, 
Bessi, Saboci". C. Miiller connects the Biessi with the Galician town of 
Biecz, I, p. 426. A still more obvious trace of them is the name of 
the Bezkydy or Bieskiden, a chain continuing the small Carpathian 
mountains towards the north. The Dacian element -bokoi re- appears in 

The Arsietai may have some connection' with the Ptolemaic town 
Arsenion east of Bohemia, due south of Kalisia (the present Kalisz in 
Poland). Cf § 23, i. 

Burgiones = Bur(i) Tabula, the alter-ego of Ptolemy's Lugoi Buroi 
in Bohemia = Kuriones in interior Germany. It is a well-known east 
Germanic tribe. The Ptolemaic Burgiones and Buroi stand fairly vis-a-vis 
and thus mutually confirm eachother's position. 


The Anartofraktoi evidently are relations of the Anartoi in Roman 

The tribes Karpianoi, Tagroi, Koistobokoi *transmontanoi may re- 
present a route leading from Karsidava (Karrodunon) down the middle 
Vistula to Askaukalis where it joins the line from Carnuntum to the 
Prussian Amber coast. It may be regarded as a continuation of the 
combined routes I, 3 and IV. 

The Karpianoi are a historically well-known tribe, homonymous 
with the Carpathian mountains. Their place roughly corresponds to that 
of the mediaeval Bielo-Chrobati or Bili-Charvati, a Slavonian tribe. As 
the Carpathian mountains were in the Old Norse Saga of Hervor called 
Harfa5a fjpll, it is probable that the "White Charvati" have inherited 
the name of their Dacian predecessors, or of the homonymous mountain. 
The Ptolemaic duplicate Harpioi seems to point towards a Gothic form 
with the same initial letter H that occurs in Old Norse and in Slavonian. 

The Tagroi are by C. Miiller, I, 431, referred to a Dacian inscrip- 
tion, found near Szent-Miklos in Hungary and containing the word "tagro- 
getzige" ("Tagro-Jazygian"?). 

The Koistobokoi *transmontanoi are the Dacians of the extreme 
north. Miillenhofif, "Deutsche Altertumskunde" II, p. 83, has transplanted 
them to northern Hungary, and also Wietersheim-Dahn in the "Geschichte 
der Volkerwanderung" and Bremer in his Ethnography place them south 
of the Carpathian mountains. This theory is based on a statement of 
Dio Cassius LIII, 12, who says that the Hastings (Astingoi), after vainly 
asking for admission into Roman Dacia, were provisionally allowed to 
leave their wives and children there while their warriors were attacking 
and conquering the region of the Koistobokoi, according to arrangement 
with the Roman governor. — It appears from Dio's words that the 
emigrated Hastings had their head-quarters south of the Carpathian 
mountains during their undertaking against the Koistobokoi, and we 
might certainly have accepted Miillenhoff's interpretation if we had not 
had the map of Ptolemy. But it is absolutely contradicted by this 
authority, and there is not the slightest reason for rejecting Ptolemy's 
map of the Dacian regions north of the Carpathian mountains: this 
section proves one of the very best parts of his work. Consequently, 
we must interpret Dio's statements quite otherwise than Miillenhoff does. 
The Hastings, a well-known branch of the Vandals, lived in Silesia. 
After being refused admission into Roman Dacia, their warriors did not 
stay south of the Carpathian mountains, but returned to Silesia, in order 
to attack their immediate neighbours, the Dacians of present Poland. — 
Our assumption is not only natural in itself, but it is also confirmed by 
two further circumstances. — i. The original map, serving as base of 
the corresponding Ptolemaic section, designated the Polish Koistobokoi as 


"transmontani", i. e. living north of the Carpathian mountains. The affix 
was intended to distinguish these Koistobokoi from their name-sakes in 
Roman Dacian (cf. Fig. 17), but Ptolemy or his predecessor misunderstood 
it, conceiving "Tranomontanoi" as a separate name, exactly as he se- 
parated the neighbouring Basternai from the synonymous Peukinoi. — 
2. The northward extension of the Dacian nationality appears from the 
Ptolemaic town Setidava, placed in Germania beyond Kalisia, i. e. north 
of the present Kalisz in Poland. This town, with the typical Dacian 
name on -dava, is evidently the outpost of the Koistobokoi transmontanoi 
towards the north-west, thus proving the extension of their territory to 
the lower Vistula. Its ethnic significance was already realised in this 
sense by Zeuss, "Die Deutschen", p. 263. 

The station north of Setidava is Askaukalis which seems to be the 
present Osielsk near Bromberg where the Vistula suddenly turns from a 
westly- direction due north-east, see the learned research of the Polish 
author J. v. Sadowski, "Die Handelsstrassen der Griechen und Romer" 
(1877), p. 58, and map. Askaukalis may be regarded as the junction of 
two mercantile roads, the one (I, 2) coming from Dacia along the upper 
Vistula, the other from Bohemia passing Kalisz. Henceforth, the amal- 
gamated routes continue until they reach the amber-producing region in 

The Jazygian towns Parka and Pession seem to be resp. the pre- 
sent Parkany near Komorn, and the present Pest. Both Ptolemaic towns, 
it is true, lie at a certain distance from the Danube, whereas the modern 
correspondences directly touch the river, and the different position of 
Pession and Pest is by C. Miiller regarded as sufficient reason for reject- 
ing the identification. We might have admitted his reasoning as plausible, 
if it concerned only one equation. But the case of Parka = Parkany is 
parallel, and if the same geographical objection is raised against both 
equations, it ceases to be an objection. As both Pession and Parka lie 
sonth-east of their modern correspondences^ the Ptolemaic localisation 
seems to betray a common displacement, originating from the Ptol. con- 
structor's wrong interpretation of a local prototype (Ad^). 

Partiskon, in Jazygia, is situated near the river Theiss which was 
in ancient times called Pathissus or Parthiscus, according to Pliny and 
Ammianus Marcellinus. 

Finally, we will draw attention to a general fact which may in 
several cases assist us in tracing the survivals of the ancient nomenclature 
on Dacian ground. 

Numerous Dacian towns or stations are homonymous with rivers or 
lakes. The same onomatic connection may occur in other parts of 


Europe, indeed, but here it seems especially frequent. And it is worth 
noticing that the invading Slavs were far less inclined to forming 
"potamic" names of settlements. E. g., the Moesian stations at the 
mouths of the rivers Isker, Vid, Osem, and Jantra have all lost their 
"potamic" names which occur on the Tab. Peuting. 

In the synopsis below, we shall register the cases concerned occurring 
on the ancient maps of Dacia. 

a. Settlements named after rivers or lakes. 

Ptol. Tibiskon, Dierna, Amutrion, Potula- (river & lake), Netindava 
(lake), Partiskon. Tab. Peuting. : Bersovia, Apo, *Cebonie. 

b. River or valley named after settlement (cf. Ogost running through 
Augustas in Moesia). 

Sar(mati)-Getias, the river of Sarmise-Getusa. Perhaps the present 
Burdea, passing Buridava. Val Fratestilor near Fratesti, the an- 
cient Frateria. 

j. Conclusion. 

To sum up, we should like to state that the analysis of the Ptole- 
maic map has shed light on ancient Dacia to an extent which could 
scarcely have been expected. If we bear in mind how little history tells 
us of Dacia during the times of the Roman dominion, the result of our 
cartographic studies may be called comparatively fruitful. 

ADDITION. Petrodava alias Patridava seems to be the present Piatra, according to 
d'Anville, "Mem. de I'Ac." XXVIII, p. 459. 



a. Summary of Contents. 

Bi & B2 are itineraries, describing the mercantile road from the 
middle Danube to the mouth of the Vistula, and containing mountains, 
rivers, tribes, and towns. The prototypes are duplicates of eachother; 
scattered duplicates occur in Acde and E. There are Latin marks; B2 
may have been translated into Greek before the stage of Ptolemy. The 
prototypes were executed after the introduction of a well-established 
amber trade under the reign of Nero (54 — 68 A. D.). Affinities with 
Strabo and Tacitus. Cf. Figures 3, 11, 12, 30, 31. 


b. Ptolemaic Localisaton. 

The Ptol. constructor has locaHsed Prot. Bi correctly within the 
northern region of the collective prototype A. We should never have 
discovered the separate existence of Bi, had we not had the alter- ego, 
Prot. B2. 

The latter prototype is displaced in westward direction, partially also 
towards the south-west. Yet the displacement does not affect the southern 
and northern limitations of the prototype, i. e. the Danube and the Baltic, 
and thus the parallel with Prot. Bi is quite easy to observe. — Prot. B2 
has enriched the Ptolemaic map of Germany with duplicates of the rivers 
Vistula and Oder, here called Svebos and Chalusos. We identify Svebos 
with the eastern frontier river of the Tacitean Swabia, i. e. the river 
Vistula. East of the Svebos, Prot. B2 places the Sidinoi, exactly as Bi 
places their alter-ego Sudinoi east of the Vistula. Chalusos runs directly 
north from the region of Kalaigia in B2, and the Oder (Viaduas) runs 
directly north from the region of Kalisia (now Kalisz) in Bi. Probably, 
Kalisia- Kalaigia was the capital of the Tacitean Helisii (read *Halisii), 
who would then have lived round the river Chalusos. We regard the 
name Viaduas as identical with Vistulas, borrowed from another proto- 
type (F}). It has certainly nothing to do with the name of the Oder, 
although geographers now unanimously assume the identification. 

A more fatal confusion was caused by another displacement of details 
from B2, due to the Ptol. constructor. The Markomanoi of B2 are on 
Ptolemy's map placed south of the mountain Sudeta of Bi, whereas the 
"tribe" Sudenoi of B2 — in reality = the Sudetes — appears south of 
the Markomanoi. Modern cartographers, in interpreting Ptolemy's map, 
erroneously regarded the Markomanoi as the "fixed point", and as this 
tribe undoubtedly occupied Bohemia, the Sudetes were consequently 
identified with the mountains north of the latter country. Nowadays the 
chimera is adopted even in popular nomenclature. The "fixed point", 
however, is not the Markomanoi, but the Ptolemaic design of mountains, y 
which clearly shows that the Sudetes lie south of Bohemia and are the 
western Bohmerwald. 

c. Definition of Limits. 

The area Bi & B2 coincides with that of Prot. F in the Baltic 
region. As both Bi 81 F are correctly localised by the Ptol. constructor, 
it is difficult to discern their elements, as soon as they do not betray 
their origin through their occurrence in duplicate series. 

Bi touches Prot. E towards the north-east, and Prot. Ac {Ae}) towards 
the south-east. There is no confusion, as Bi remains within Germanic 


territory (apart from Sudinoi = Sidinoi B2), whereas the latter two 
prototypes are by the Ptol. constructor limited to Sarmatia and Dacia. 
The correctness of Bi sharply contrasts Ptolemy's completely displaced 
localisation of E. 

82 on its western side touches the Prototypes Aa, Ab, C, and D. 
There seems to be no serious confusion. — Ptolemy's wrong localisation 
of B2 contrasts his correct localisation of Aa and Ab. On the other 
hand, the displacement of B2 contrasts the opposite displacement of C. 
B2 has been pushed towards the west, and C towards the east, with the 
result that the *Buriones and Marvingoi B2 from eastern Germany collide 
with the *Chattvaroi C from the mouth of the Rhine. The tribes of D 
distinguish themselves through the addition of "Sveboi". 

d. General Topographic Scheme. 

The presence of mountains in Bi and B2 appears from the duplicates 
Asbikurgion — Bikurgion, Sudeta-r-Sudenoi. But it may perhaps not be 
taken for granted that the entire Ptol. design of Bohemian mountains 
belongs to Bi. We have assumed a collective oro- and hydrographic 
map of Germany and Bohemia, viz. A, into which Bi could be intro- 
duced as a supplement. 

It is more self-evident that both Bi and B2 contained two rivers, viz. 
the Vistula and the Oder. Their arrangement on the Ptol. map still 
preserves an obvious parallelism. Prot. Bi seems to have contained a 
third river, which starts from the mountain Askiburgion and is supposed 
to join the Vistula after passing directly west of the town Kalisia. It 
may be the present Prosna which, after passing directly west of Kalisz, 
joins the Warta, — not the Vistula. But it may, perhaps, also be an 
original road-line, misunderstood by the Ptol. constructor. 

The entire so-called "Vistula" Bi between its source and the town 
Askaukalis is in reality no river, but a road-line, leading from the source 
of the Vistula to the large turning of this river near Bromberg or Osielsk. 

The Ptol. river Vistula, apart from representing in its superior course 
an original road-line, forms the frontier between the Ptol. sections Ger- 
mania and Sarmatia from its mouth to its source. The frontier continues 
farther south without following any physical line on the Ptol. map, till it 
reaches the Sarmatian mountains: in this interval it would have been 
correct to make the frontier follow up the Vistula which, as a matter of 
fact, starts from the said mountain complexe. We may take it for granted 
that the piece of frontier without physical underlining reflects the con- 
tinuation of the road-line on the original map Bi. The existence of an 
itinerary leading from the Sarmatian mountains to the mouth of the 
Vistula evidently influenced the Ptol. scheme of map division in a funda- 



mental manner: this pronounced line was used as mark of distinction 
between the sections Germania and Sarmatia. As a matter of fact, the 
road fairly coincided with the demarkation of the main nationalities. Only 
few Gothonic tribes were situated east of the road, such as *Buriones 
(Burgiones), Basternai, and Gythones, whereas only a single Dacian town 
appears on its western side, viz. Setidava. 

The itinerary Bi was of fundamental importance, not only as a means 
of distinguishing the sections Germania and Sarmatia, but also from an- 
other point of view: its stations were used by the Ptol. constructor as 
marks of astronomic orientation. This fact will appear from the following 
list of correspondences. 

Actual latitude 

Mouth of the Vistula 

Askaukalis, near Bromberg (Osielsk ?) . . 

Kalisia = Kalisz 

Mountain Askiburgion, south-eastern 
extremity, = Jesenik 

Sarmatian mountains, northern ex- 
tremity, = Bieskiden 

Eburo(duno)n = Brno, Briinn . 

Sarmatian mountains, southern extremity, 
near Pressburg 

Danube, curve at Kurta, near the 
present Raab 

Mouth of the Vistula 

Eburo(duno)n = Brno, Briinn 



c. iVs 


} c. 1V6 

I'/ 4 








c. V^ 

C. I 


Actual longitude 
(from Greenwich) 



c. 3 

io6 Ptolemy's maps of northern Europe 

The Ptol. measurements, it is true, are generally not quite exact, but 
we could scarcely have expected them to be better in a country like 
Germania. • 

It is well-known that Ptolemy describes in the preface how the longi- 
tudes and latitudes throughout his work have been calculated by com- 
paring the statements of numerous maps and travellers, — an operation 
which he calls extremely difficult. If we would take these words literally 
in interpreting the Ptol. towns in Germania, as numerous previous 
scholars have done, we should certainly in most cases be mistaken, for 
the vast majority of the localisations are merely fictitious. But it is 
different with the area of Prot. Bi: here we have really a sample of 
those itineraries which served as foundations of the astronomic calcula- 
tions in the trustworthy parts of Ptolemy's work. Its position is quite 
solitary in the middle and northern parts of the Ptol. Europe and may 
be regarded as a most prominent feature of its literary individuality. 

e. Statistical Features. 

The Prototypes Bi & B2 seem to contain a fairly equal selection 
of the most usual geographical categories: rivers, mountains, tribes, and 

They thus contrast the prototypes Aa, Acde, and E, which seem to 
have recorded mainly tribes in the neighbouring regions. 

A different contrast is represented by Ab which contains no tribes. 

Bi & B2 betray no sure traces of the "ethno- topic denomination" 
which characterizes Prot. F, e. g. Venedai with Venedian gulf and moun- 
tain, Peukinoi with mountain Peuke, etc. 

The comprehensive statistical selection within Prot. Bi & B2 corre- 
sponds to the importance of the mercantile road to the amber coast. It 
is moreover emphasized by the fact that the region concerned shows the 
highest percentage of second class towns in the whole of Germania out- 
side the Roman territory. 

The following synopsis illustrates the distribution, as it appears in 
four of the oldest MS. atlases. 

















in the territory of the moun- 

taneering Celts. 
Duplicate name. Astronomic 


Brunn or Brno, 
capital of 













Susudana ] 

Setidaua J 







2 (1?) 













2 ' 





in the territory of the moun- 

taneering Celts 
in the territory of the Helisii 

Duplicate: Kalisia 
in the territory of the Koi- 

Duplicate name 
(Duplicate: Astouia Alisos) 



Kalisz, capital of 










Komorn ? 








3 . 


1 (sic) 
3 . 







Fictitious town 
Duplicate Arsikva 

V. 1 



(in the territory of the Ar- 

Duplicate: Aregelia 

Our survey seems to confirm the statistical scheme of the MS. atlases, 
so far as it is possible to speak of verification concerning regions which 
have been almost completely revolutionized during the age of great 

The town of the first class, Eburodunon, till this day is the most 
important of those which have survived. 

Among the towns of the second class, Kalisia is the most remarkable. 
It seems to be the capital of the Tacitean Helisii, and it survives till 
this day as Kalisz, the capital of a homonymous government. 

5 of the 8 towns among our instances of the second class belong to 

io8 Ptolemy's maps of northern Europe 

eastern Bohemia or the regions south of it, — an area which was already 
at the beginning of our era well known by the Romans. It is the region 
near the capital Eburodunon. 

The third class, finally, is accompanied by no additional classical 
evidences and affords no sure modern survivals. Redintuinon, Nomisterion, 
and Arsenion belong to comparatively remote districts, and Marobudon 
is fictitious, constructed on the base of a Tacitean passage mentioning 
the castle of king Marbod. The entire class, consequently, seems to be 
of inferior importance. 

f. Occurrence of ^Duplicates. 

The area of the duplicate series Bi & B2 covers the provinces of 
Oesterreich (Rakousko), Moravia, Bohemia, Thuringia (partially), Silesia. 
dBu^ica (Lausitz), Posen, eastern Pomerania, and Prussia. 

The following names re- appear in other prototypes. 

Rakatriai Bi^ Rakatai B2 = Ratakensioi of an interpolated prototype in 

Dacia (^?). 
*Kotnoi Bi, *Koteinoi B2 = Kotensioi (Kontekoi) of the mentioned 

Dacian prototype. 
Buroi Bi, *Buriones B2 = Burgiones Acde (= (Ouis)burgioi?). 

^. Linguistic Marks. 

Latinisms. Plural on -z: Lug2-(-dunon) B2. Ablative form -one: 
Singone Bi. ng, nk\ Marui;2^oi B2, Si;/^one, Asa;//^a Bi. 

Apart from these, there are some indications which seem to suggest 
that the two prototypes had been translated into Greek, before the Ptol. 
constructor combined them with Prot. A. They contain some mis- 
readings which are best explained by the assumption of a Greek original. 

POYnKAIOI Bi < *POYnKAIOI. (Does this name originate from 

Prot. F}), 
KOrNOI Bi < *K0rN0r. 
KAAAiriA B2 < *KAA^/CIA. 

In all these cases, Latin letters would not so easily cause the same 
misreadings. Finally, we observe the Greek word for "grove" : Limios 
alsos, contrasting the Latin words used in the sphere of Prot. A. 


h. Literary Milieu. 

The southern sphere of the prototypes Bi & B2 was well known to 
the Romans, owing to their constant interference with the affairs of the 
Marcomans and Quades. Cf. the rivers Duria and Marus, mentioned by 
Pliny IV, c. 81, in the frontier districts of Vannius, king of the Quades. 
Marus is the present Morava or March from which the district of Moravia 
draws its name. Duria seems to be the present Thaya, in Cechian 
called Dyje. Tacitus mentions the river Cusus (now Gusen), "Ann." II, ch. 63. 

The Baltic regions were explored considerably later. 

Agrippa had some ideas about them, but only vague. He says that 
Dacia is limited by the Ocean in the north, and by the river Vistula in 
the west. The dimensions of the country are given thus: CCLXXX 
milia passum in the longitude, CCCLXXXVI in the latitude. 

The extent of the area of Dacia towards the north quoted above 
agrees with Prot. Bi & B2 which place the Dacian town Setidava in the 
coast region of the Baltic Ocean, and due west of the river Vistula, 
as an isolated outpost of the Dacians among Germanic surroundings. 
Cf. the presence of the Dacian Koistobokoi on the opposite side of the 
Vistula, according to Prot. Acde. 

Apart from that, Agrippa seems to have had no information about 
Baltic regions. 

The first more detailed observations represented by Bi & B2 date 
from the times of King Marbod's great Swabian Empire, which embraced 
even the *Sudines (Sibinoi) in Prussia, cf Strabo VII, p. 291. 

The intercourse between Rome and the regions about the mouth of 
the Vistula was increased under the Emperor Nero, when a regular 
amber trade was established, cf. PUny XXXVII, ch. 45. 

If we examine the prototypes Bi & B2, we shall find the older stage 
of Roman topographic knowledge expre;ssed by affinities with Strabo, 
whereas the later increase of commercial intercourse appears from the 
numerous affinities with Tacitus. 

Affinities with Strabo and Tacitus (VII, p. 290 seq., "Germ." 
ch. 42 seq.). 

Lugoi Bi (Lugi B2) = Luioi Strabo, Lygii Tacitus. 

Omanoi Bi = Atmonoi Strabo, Manimi Tacitus. 

*Rugiklioi Bi with town Rugion = Mugilones Strabo, Rugii Tacitus. 

Affinities with Strabo. 

Sidones Bi = Sidones, a branch of the Basternes. VII, p. 306. 
Omanoi Bi, more related with Strabo's Atmonoi ibd., than with the 

Tacitean Manimi. 
Sudinoi Bi, Sidinoi B2 = Sibinoi Strabo. 


Affinities with Tacitus. 

Varistoi Bi = Varisti **Germ." ch. 42. 

*Kotnoi Bi, *Koteinoi B2, with Celtic town-names = Coteni *'Germ." 
ch. 43, with Celtic language ("Gallica lingua"). 

Ironworks east of Bohemia Bi =. ironworks of the Coteni, ibd. 

Mountain Askiburgion, dividing the Lugoi Bi = "a continuous mountain 
chain divides Swabia" (''dirimit scinditque Suebiam continuum montium 
iugum") "Germ." ch. 43. 

Division of the Lugoi in several tribes Bi & B2 = "the Lygian nation 
is the most extended, and divided into several tribes" ("latissime 
patet Lygiorum nomen, in plures civitates diffusum") ibd. 

Lugoi Buroi Bi, *Buriones B2 (south of Askiburgion) = Burgiones Acde 
= Buri "Germ." ch. 43 (evidently south of the "continuum jugum"). 

Marvingoi (beside *Buriones) B2 — Marsigni (beside Buri) ibd. 

Lugoi Omanoi Bi = Lygii Manimi ibd. (contrasting Strabo's Atmonoi, 
who are represented as a branch of the Basternes). 

Kalisia Bi, Kalaigia B2, near the river Chalusos B2, cf. Helisii "Germ." 
ch. 43. 

"Grove of Limis" = the grove of the Nahanarvali, a Lygian tribe ibd. 

It might be tempting to add Eluaiones = Helvaeones ibd. But as the 
name seems to re-appear in Prot. E as Igylliones, it would rather be- 
long to the duplicate prototype F. It is, however, not excluded, that 
the name Eluaiones occurred both in Bi and 'F. 

Taking it as a whole, it must be admitted that the affinity between 
the prototypes Bi & B2 and Tacitus is striking. 

i. Examination of Details. 

In spite of partial displacements, the parallel between the two 
duplicate series Bi & B2 remains easy to trace. Only in few cases, the 
order of links is disturbed. See our figure 19, which speaks for itself. 
It is very fortunate that the duplicate series exist, for several of the 
doubled names are preserved nowhere else, be it in modern topography 
or in the ancient. 

We shall now regard the single names, comparing them with the 
evidences of mediaeval and modern geography. 

I. Rakatriai Bi, Rakatai B2 = Ratakensioi on the Ptol. map of 
Dacia beside Kotensioi, cf. *Kotenoi, neighbours of the Rakatriai. Rakousko 
is the Cechian name of Austria, borrowed from the province of Nieder 
Oesterreich. A mediaeval castle of that province, called Rakoutz, is 


supposed to be the present Raabs. Cf. Safarik, "Slavische Altertiimer" 
I, 50 seq., Miillenhoff, "Deutsche Altertumskunde", II, 331. 

2. Singone Bi is the Latin ablative form of a name that seems to 
be Dacian, cf. Singidava in Dacia, and the Daco-Celtic town Singidunon 
in Moesia. The Latin flexion betrays that the station was well known 
by the merchants. 

3. Eburon Bi, Eburodunon B2, corresponds to Brno or Briinn, the 
capital of Moravia. The Ptol. distance of Eburon from the Danube, like 
that of Briinn, is exactly one degree of longitude. Eburodunon belongs 
to the points of astronomic observation recorded by Ptolemy in Book 
VIII, VI, 3, and is consequently decorated with towers on the map, but 
the resulting localisation is too near the Danube for Briinn; the sur- 
rounding names from B2, such as Baimoi and Arsikva, show the same 
dislocation towards the south. The present forms Brno and Briinn, with 
loss of initial £, may remount to the Celtic accentuation which also 
appears in the French forms of the same name : Embrun in south-eastern 
France, and Iverdon in Switzerland (Germ. Ifferten), both with the stress 
on the last syllable. 

4. *Arsekvia Bi, Arsikva B2, is probably a town of the Dacian 
tribe of Arsietai, placed by Ptolemy in the directly contiguous part of 
Sarmatia. The place Arsenion Bi in the neigbourhood =• Ar(e)gelia, 
Aregeouia B2, also seems to belong to them. 

5. Sudeta ore Bi, "tribe" Sudenoi B2, = the western Bohmerwald. 
The Sudeta ore are placed south of the Bainochaimai = Bohemians, and 
the Sudenoi south of the Markomanoi, also = Bohemians. 

6. Bainochaimai Bi, Baimoi B2, = Bohemians. The vocalisation at 
in *Baio- is a sign of enlarged local experience, as the preceding classical 
authors write constantly oe or ot, owing to the connection with the well 
known Celtic tribe of Boji. 

7. Varistoi Bi. The well known tribe of Varisti, later occupying 
the "pagus Varascus" in Burgundy, according to its own national tradi- 
tions originated from the district of Stadewanga near the river Regen, 
i. e. near the present Regensburg. See Egilbert's "Vita S. Ermenfredi", 
Acta Sanctorum Vol. VII, Sept. 25. The localisation agrees with the 

8. *Kotenoi Bi, *Koteinoi B2 = the Kotensioi (Kontekoi), errone- 
ously placed on the Ptol. map of Dacia (from B}). It is a well-known 
tribe of mountaneering Celts in Bohemia (Tacitus). The ironworks 
(siderorycheia) on the Ptolemaic map are placed in their neighbourhood. 
According to Strabo, the silver mines of Sisapon in Spain were called 
"Kotinai". As the Celts were the pioneers of mountaneering in most 
parts of Europe, "kotinai" seems to be the Celtic word for "mines", 
and Kotenoi would be "miners". In Cechian, kutati is "to mine, to dig", 


kutny = "mining", and an important mining centre in the region of the 
Kotenoi is called Kutna hora, Germ. Kuttenberg. The Slavs certainly 
learned the mining technique from the Kotenoi, and so probably also 
adopted its Celtic terms. Consequently, the name Kutna hora may more 
or less directly remind the Celtic tribe of Kotenoi. 

9. Sidones Bi^ known from Strabo as a branch of the Basternai. 
On the Ptol. map of Sarmatia, the Basternai are placed fairly vis-a-vis. 

10. Lugoi Buroi Bi^ *Buriones B2 = Burgiones Acde^ placed in 
Sarmatia fairly opposite the Buroi. The Buri are well-known from Ta- 
citus and other classical authors. 

11. Marvingoi beside *Buriones B2 = Marsigni beside Buri, Tacitus. 
The Marvingoi may have some connection with Maurunga, a mediaeval 
name of the regions east of the Elbe, = the epical Mornaland in the 
Old Norse poem of Oddrunargratr (land of the With-Myrgingas in Wid- 
sith?). The mediaeval name, later assigned to the Slavs, was in the 
"Chronicon imperatorum et pontificum bavaricum", MG. SS. XXIV, 222, 
changed into Mauritani. The linguistic connection with Marvingoi is 
not normal, but accidental coincidence is on the other hand also un- 

12. Korkontoi Bi, are the inhabitants of the Krkonosc, or Riesen- 
gebirge. Cf. Safarik, '^Slav. Altert." I, p. 486. Mullenhoff, "Deutsche 
Altertumskunde" II, p. 373, rejects the equation, because it does not 
satisfy the strict laws of phonetic correspondence. His objection, how- 
ever, is not justified, as important local names are often subjected to 
arbitrary transformations, owing to popular fancy etc. 

13. Mountain Askiburgion (in numerous MS. atlases, e. g. Urb. 82: 
Asbikurgion) Bi, "town" Bikurgion B2. This chain is generally identified 
with the present Jesenik (Germ. Gesenke), as both names signify "Ash- 
mountain". Perhaps, Askiburgion might also be reflected by the present 
Je§ted or Jeschken in northern Bohemia. The position would agree well 
with the north-western extremity of the Askiburgion. It is not excluded 
that the original name might have been developed or translated differently 
in the local dialects. 

14. ' Teuriochaimai Bi (Turonoi B2}).- The so-called tribal name is 
derived from the name of a district which may signify the "Home of 

15. Arsenion Bi, Argelia, Aregelia, Aregeouia B2, on the northern 
frontier of Bohemia, according to Bi. Probably a frontier town of the 
Arsietai in independent Dacia, cf. under 4. 

16. Kalisia Bi, Kalaigia B2., in Bi directly east of the river (?) 
corresponding to the present Prosna*, in B2 south of the river Chalusos. 
Probably the capital of the Tacitean tribe Helisii, the epical Haelsingas 
who are mentioned in the poem of Widsith. It is the present Kalisz, 

§ 2 3- LOCAL PROTOTYPES Bi & B2 I 13 

the capital of a homonymous government. The present K of the name 
may remount to Dacian pronounciation. Ptolemy places Kalisia on 52,50 
of latitude, whereas the actual position of Kalisz is 51,47. The Ptol. 
distance from the mouth of the Vistula is 3,10, whereas the real is about 
2^/2. In both cases, the difference is of little import. The Polish 
scholar J. v. Sadowski points out that Kalisz occupies a position on the 
most convenient route leading to the ford near Konin between the moors 
of the Warta, s. "Die Handelsstrassen der Griechen und Romer durch 
das Flussgebiet der Oder, Weichsel", p. 57. 

17. Lugoi (Dunoi) Bi, Lugi-(-dunon) B2. Inhabitants of the present 
^u^ica or Lausitz, a well-known eastern Germanic tribe. 

18. (Lugoi) Dunoi Bi, (Lugi-) -Dunon B2. Inhabitants of the epical 
Dun-hei5i, or "Dun-heath", mentioned in the Old Norse poem "Battle of 
Huns" (Hervararsaga) on the frontier against Hunland, i. e. Hungary, 

19. Siliggai B2. The present Silesians, Pol. Slezani. The Slavonic 
form is developed normally from a Gothonic Siling, exactly as Slav, 
knez < kuning, "king". As a branch of the Vandals, the Silingians 
played a great role during the migration age. 

20. Limios alsos, "grove of Limis", B2. Probably identical with the 
sacred grove of the Lygian tribe of Nahanarvali, mentioned by Tacitus. 
Cf. C. Miiller I, p. 270. 

21. Lugoi Omanoi Bi. The Lygii Manimi of Tacitus; the Atmonoi 
of Strabo, represented by him as branch of the Basternes. 

22. Setidava Bi, Susudana B2 (Cod. Vatic. 191). A town with the 
well-known Dacian element -dava. Its presence in these northern regions 
of Germany, not far from the mouth of the Vistula, is supported by the 
Ptolemaic localisation of the Koistobokoi *transmontanoi, who are placed 
on the opposite side of the Vistula. These northern Koistobokoi were a 
great independent Dacian tribe: they fought against Rome in the Mar- 
comannian war (Julius Capitolinus ch. XXII), ravaged Greece (Pausanias 
IX, 34), were defeated by the Vandalian tribe of Hasdings, but revenged 
by the Dankriges (Dio Cassius, LXXI, 12). 

23. Askaukalis Bi, probably = Astouia & Alisos B2, perhaps "the 
town Astouia of the tribe called *Halisii", cf. the name Sarmize — Getusa, 
signifying the mixture of two ' nationalities. The town concerned is the 
last station on the route and must consequently have occupied an im- 
portant position. The German scholar Voigt has proposed to identify it 
with the present Osielsk near Bromberg, and Sadowski accepts this sug- 
gestion as strikingly convincing. As the Ptol. spelling of Askaukalis 
is all but certain, nothing prevents us from assuming that it might 
be continued in the form of Osielsk. Still more decisive is the topo- 
graphical argument: Osielsk lies exactly at the point where the Vistula, 
after its large curve through Poland, suddenly turns from sharp west- 



ward direction towards the north-east. Here the mercantile road from 
the Danube, after leaving the Vistula in upper Silesia, again joins the 
river in order to follow it to its mouth; such place certainly demands 
a station. The distance of Askaukalis from Kalisia is about 1V2 degree 
of latitude, whereas the distance of Osielsk from Kalisz is about 1V4. 
Thus the Ptol.' localisation seems well verified also from the astronomic 
point of view. 

24. *Rugiklioi with town Rugion at the Baltic coast = the well- 
known Gothonic tribe of Rugi, the epical (H)ulme-Rugi of Jordanis, the 
Holm-Ryge of Widsith. It is only not necessary that the names men- 
tioned belonged to Prot. Bi\ they might also have belonged to Prot. F. 
The Rugi are mentioned by Tacitus as the most northern of the tribes 
in eastern Germany, a fact that makes us inclined to refer them to B/, 
owing to the close affinity between this prototype and Tacitus. 

25. Vistulas Bi, Suebos B2. The river Vistula, the eastern frontier 
of the Swabians, according to the Strabonian and Tacitean description. 

26. Sudinoi Bj, Sidinoi B2. The mediaeval Sudovitae, a Prussian 
tribe, inhabiting the present district of Sudauen. 

27. Galindai Bi (or Prot. F}). The mediaeval Galinditae, another 
Prussian tribe. 

j. Conclusion. 

As result of our comparison, the topography of Bi and B2 may be 
called well verified. 

These twin prototypes, like Ad & Ae, supply a valuable piece ot 
topography and ethnography from a region, which lost most part of its 
ancient population and nomenclature during the age of migration. Their 
evidences enable us to trace exactly the localisations of different 
nationalities along the route of Roman amber trade from the Danube to 
the Baltic, viz. Pannonians, Celts, Dacians, Gothons, and Lithuanians. 
In § 22, we have pointed out the importance of the town Setidava Bi 
= Susudana B2, as an outpost of Dacian nationality in northern regions 
which are as a rule wrongly attributed to «the Gothons. 

ADDITION. R. Much, "Die Stadte in der Germania des Ptolemaus" ("Zeitschrift fflr 
deutsches Altertum" XLI, 97. 1897) already sets forth a long series of those critical obser- 
vations which we have made above concerning the Ptol. misreadings and wrong localisations. 
In other points, his assumptions would lead to different results. Astouia, var. Aistouia, is 
interpreted as a Latin "aestiva sc. castra", cf. Velleius II, 117: "mediam ingressus Ger- 
maniam .... trahebat aestiva". Alisos, Lakiburgion, Budorgis-Budorigon are identified with 
the Rhenish towns Alison, Askiburgion, Budoris, and Susudana-Setidava with Zusidava in 
Dacia. If the suggestions concerning the first-mentioned five towns be correct, it would 


imply the assumption of at least one additional prototype. We do not think that the 
existence of a Dacian town Setidava in eastern Germania need be rejected, as the Ptol. 
Koistobokoi *transmonianoi prove the presence of Dacians in Poland (cf. p. 113, 22). Our 
main results concerning the prototypes Bi & B2 do not seem to be affected by Much's 
divergent statements. 


a. Summary of Contents. 

Prot. C is an itinerary, describing north-western Gaul, Belgium and a 
part of north-western Germany, containing rivers, tribes and towns. 
Duplicates occur in Aa. The prototype has Latin marks, but was per- 
haps translated into Greek before the stage of Ptolemy. There is close 
affinity with the Itinerarium Antonini and the Tabula Peutingeriana. 

Cf. Figures 1,21, 22, 23. 

b. Ptolemaic Localisation, 

The Ptol. constructor has introduced Prot. C into the corresponding 
parts of Prot. A in such a manner that C is absorbed without leaving 
directly visible traces, so far as physical outlines are concerned. Yet 
the presence of C is apparent from the eastward displacement of the 
accompanying names, especially the duplicates. Most of the towns con- 
cerned have been noticed by C. Miiller. 

At the outset, it is not obvious whether all of the displacements' 
must be regarded as betraying Prot. C, or whether some of the names 
concerned might be derived from other sources. Provisionally leaving 
this question undecided, we shall register any cases of displacement ob- 
served by us in Gaul, Belgium, and north-western Germany. 

We begin with western Gaul. 

Redones, the people of the present town Rennes, form the starting 
point of the displacement, being removed from the region of the lower 
Loire to the middle course of that river. In Prot. A, the Ptol. constructor 
must have found both the Redones and their town Kondate missing, but 
he found a name-sake of the latter on the middle Loire, — both towns 
are recorded by the Tabula Peutingeriana — , and consequently he pushed 
the Redones thither. Once begun, the displacement continued, as we shall 
see by regarding the position of their neighbours. 

Namnetai, the people of Nantes, emigrate from the mouth of the 
Loire to the mouth of the Seine; their town Kondeouinkon, now Nantes, 


Ii6 Ptolemy's maps of northern Europe 

being likewise removed. Their correct place is still marked by their 
mutilated alter-ego Samnitai, originating from Prot. A. 

Abrinkatuoi, the people of Avranches, from western Normandy to the 
mouth of the Seine, with their town Ingena, now Avranches. 

Ratomagos, now Rouen, from the Seine towards the east. The 
duplicate of this town, originating from Prot. A, is in return pushed a 
little west of the river. 

A third duplicate, betraying possibly the contrast between Prot. C 
and A, is roMorinoi = Morinoi, in the present Flanders. Accidentally, 
no displacement has occurred here worth mentioning. 

We now enter the Belgian district called Germania, mentioned e. g. 
in the Itin. Antonini. The Ptol. constructor has mistaken this whole 
district for the Germania megale of his Prot. A, i. e. the present Ger- 

The western frontier of the Belgian Germania is mistaken for the 
Rhine A which forms the western frontier of Germania megale. The 
middle course of the actual Rhine in return is mistaken for the Abnoba 
of Prot. A (A). 

In "The Scottish Geographical Magazine", vol. XXX, p. 70, we have 
suggested that the continuation of the Rhine is concealed by the moun- 
tain Melibokos and the river Weser in A. Further considerations have 
caused us to withdraw this suggestion. 

The mountain Melibokos seems to lead us too far south of Askalingion, 
which marks the place of the Rhenish town Askiburgion or Asberg. 
And Leufana = Levefano Tab. Peuting., lies at a considerable distance 
east of the Weser, whereas it ought to lie on the western border, if 
this river were to be regarded as the original Rhine of C. As a matter 
of fact, the Ptolemaic map of Germany seems to contain no physical line 
which could have been identified with the lower Rhine of C. It is per- 
haps not excluded that Prot. Aa contained a line representing the frontier 
of the Roman territory in northern Germany between the years 9 A. D. 
and 47 A. D. Such a line might have crossed the Weser and touched 
the Elbe exactly at the places where the Ptolemaic map puts the names 
Askalingion and Leufana. And the Ptol. constructor would have identified 
the lower Rhine C with this frontier line of Prot. Aa, whereas the final 
edition of the atlas eliminated the frontier line, because the Romans had 
in 47 A. D. given up their dominion over the North German coast. 

If the reasons of this displacement on German ground remain some- 
what obscure from the physical point of view, its presence is no less 
certain, as the reader will notice from the following lists of correspon- 
dences : 


Tab. Peuting. 

1234 5 67 89 10 11 

A.spingium Tab[u]lis Flenio Foro Adriani River Anatius Matilone Albanianis Leuefano Caruobe Asciburgio Nouesio. 

Prot. C. 

1235 4 76 89 10 11 

Lskiburgion Nabalia Fleum F. Abiranon River Amisias Marionis Albis(*amnis) Leufana Koinoenon Askalingion Nouaision. 

Itin. Antonini. 





Prot. C. 









We learn from these lists that the Ptolemaic names of towns and 
rivers in north-western Germany re- appear often as Belgian on the Tabula 
Feutingeriana or in the Itinerarium Antonini, either with almost identical 
forms, or under a slight disguise. 

The following easily identified towns are localised by the Ptol. con- 
structor without any traceable assonances on German ground; most of 
the equations have been suggested by C. Miiller: Leufana, Levefano 
Tab Peuting., according to C. Miiller the present Levenstein; Askal- 
ingion, Asciburgio Tab. Peuting., the present Asberg on the Rhine; 
Nouaision, Novesio Tab. Peuting. and Itin., the present Neuss on the 
Rhine; Teuderion, Theudurum Itin,, the present Tiiddern between the 
Meuse and the Rhine; Mediolanion, Mediolano Itin., perhaps the present 
Moyland near Asberg; Bogadion, Bacaco Tab. Peuting., Bagacum Itin., 
the present Bavay on the Sambre^); we may add Tekelia, mentioned by 
no other classical evidences = the present island of Texel. 

In the following cases, Belgian names of Prot. C have been absorbed 
by correctly localised German names of Prot. Aa, owing to treacherous 
assonances. Although incomplete, the assonances are sufficiently "self- 
evident" in order to deceive a Ptol. constructor, after all we know about 
his philological capacity. As a matter of fact the order of names on 
the Tab. Peuting. corresponds so well to the assonances on the Ptolemaic 
map that we cannot wonder he was mistaken. Caspingium — perhaps 
written with indistinct initial — became Askiburgion, now Asberg on the 
Rhine; Tabulis, *Nabulis > Nabalia on a homonymous river, mentioned 
by Tacitus near the Zuider Sea; Flenio > Fleum, on the Vlie Strom; 

*) Miiller suggests an identification with Burginatium of the Tab. Peuting. and Itin. Ant., 
but the assonance seems too feeble. 


*F. Adrianum, Fabiranon >> Fabaria, a Roman name of the island Bor- 
kum; river Anatius > river Amisias, now Ems; Albanianis, the present 
Alfen, near Leyden, misunderstood as *Albis amnis > Albis, the Elbe. 
We may add Orolaunum, the present Arlon (Flemish Aarlen), west ot 
Luxemburg, misread by the Ptol. constructor as Ar-taunon, and localised 
near the mountain Taunus. 

After the towns and rivers, we shall consider some tribes from the 
district Germania west of the Rhine, transplanted by the Ptol. constructor 
to Germany. 

The Tenkeroi of Prot. Aa seem to have absorbed the *Tungri of C 
= Tongri of Itin. Antonini. Inkriones, between Rhine and Abnoba, look 
enigmatic. In "The Scott. Geogr. Mag.", vol. XXX, p. 70, we have sug- 
gested an equation with the Belgo-German tribe Eburones, as the termi- 
nation -ones is very rare among the tribes of Belgium (other instances: 
Ceutrones, Olibriones)^). Intouergoi, south of the Inkriones, are another 
tribe with an extraordinary kind af name. In our above-mentioned 
research, we have identified them with Strabo's Trevagroi = Treveri, the 
inhabitants of Trier. But an examination of the Tab. Peuting. supplies a 
more plausible equation: Intouergoi = Nitiobroges. The latter name is 
corrupted by the author of the Tab. Peuting. owing to erroneous identi- 
fication with the well-known Nitiobriges near the Garonne. The second 
element -obriges Tab. Peuting. seems to be derived from Obringa, the 
Ptolemaic name of the river Mosel, and the (Niti) obriges thus would be 
connected with the 01-ibriones of Jordanis XXXVI. and the Al- obrites 
or Al -obroges of the Anon. Ravennas, IV, 24 and 26, cf. Zeuss, p. 578, 
579. The first syllable int = nit would have been more correctly spelt 
by Ptolemy, and the spelling verg. instead of brig may represent the 
vulgar Latin pronunciation, cf Borvetomagus, *Vorbetomagus instead of 
Borbetomagus, now Worms. 

Vargiones south of Intouergoi are, of course, the well-known German 
tribe Vangiones, correctly localised west of the Rhine on Ptolemy's map 
of Gaul. Perhaps they are concealed by the badly corrupted name "Rer- 
viges" beside Nitiobroges on the Tab. Peuting. 

Karitnoi south of the Vargiones = Parisi Tab. Peuting., erroneously 
identified with the well-known inhabitants of Paris. They are mentioned 
by Caesar as Caeresi, by Tacitus as Caeracates, and lived in the mediaeval 
district Pagus Caroascus north-west of the Mosel. The derivation has 
evidently been somewhat fluctuating. 

All the tribes mentioned indubitably belong to Prot. C. Continuing 

*) The reading Nikriones of one MS. {A) has by some scholars been combined with the 
Nicretes of a Roman inscription, but it is too isolated. 


farther east, we enter that region which we have in our provisional 
sketch assigned to D. 

The tribes concerned belong to the country east of the Rhine. 
Kamauoi and Chairusikoi, near Leufana, at the utmost edge of the area 
with eastward displacement, correspond to the Chamavi and Chrepstini 
on the Tab. Peuting., not far from Leuefano, at the utmost north-western 
edge of Germania. The displacement of the Kamauoi is very strong: 
from the Rhine to the Elbe. The neighbouring tribes Chattai and Tu- 
bantoi have equally been transplanted from the Rhenish districts to interior 
Germany. Kalukones = Kathylkoi Strabo: *Kauklones, or smaller Chauks, 
occupy both sides of the Elbe, according to Ptolemy's text, although 
they ought to stand in reality west of the Weser, as their alter-ego does 
in Prot. Aa. 

The presence of a tribe *Angrivarii in C may be "conjectured from 
the absurdly displaced Ptolemaic Sueboi Aggeiloi belonging to Prot. D. 
It is scarcely conceivable how it could occur to the Ptol. constructor's 
mind to place the Angles in interior Germany, if he had not been misled 
by some assonance. The form Angrivarii may have been abbreviated 
into Angri, so that only the two first syllables were legible. It is not 
excluded that the corrupted forms *'Vapi. varii" on the Tab. Peuting. 
might conceal the name of the same tribe.v The Brukteroi may also 
have occurred in C, corresponding to the Burcturi on the Tab. Peuting. 
If so, the "smaller Brukteroi" near the Rhine were really meant, whereas 
the Ptol. constructor identified them with the "greater Brukteroi" of 
Prot. Aa farther east. 

Next to the Brukteroi, we notice the Kasvaroi and the *Chattvaroi 
(Chaitvoroi), two ^tribes that ought to stand near the lov^er Rhine. ("Haci. 
Vapi. Varii" on the Tab. Peuting.??). The Ptol. constructor has trans- 
planted them east of the Abnoba; the *Chattvaroi were probably assi- 
milated with the Raetovarii, a Danubian tribe in the present district of 
Ries, mentioned in the "Notitia Dignitatum". 

Finally, the Uispoi follow, = Usipi. This is the only one of the 
displaced tribes that has retained its position near the Rhine. In return 
it has been pushed far south from the region north of Mayence to the 
slopes of the Schwarzwald. 

c. Definition of Limits. 

We may expect that the definition of limits will here cause some 
difficulties, because C is neither accompanied by a duplicate prototype 
nor limited by distinct natural or political boundaries. As a matter of 
fact, we have altered our views considerably, since we published our first 


Ptolemaic eassay in the "Saga Book of the Viking Society", vol. VIII 
(191 3), and in "The Scottish Geographical Magazine", vol. XXX (1914). 

One question concerns the distinction of prototypes in Gaul. 

As the reader will notice from our Fig. 21, the Ptolemaic displace- 
ment affects two sections which are distinctly separated from eachother. 
The one represents a region in western Gaul, which is transplanted to 
the borders of the river Seine. The other is the Belgian district Ger- 
mania which is transplanted east of the Rhine, whereas the adjoining parts 
of Germany are pushed farther east. 

Owing to the complete separation of the two displaced sections, it 
might seem questionable whether they originate from a single prototype 
or from two. As the displacement is so constantly eastward, however, 
the assumption of a single prototype seems most likely. Moreover, a 
correct map shows no chasm between the sections concerned, as the 
reader will notice by regarding our Fig. 23. 

In § 24, we shall supply further material showing that the Ptol. con- 
structor sometimes indubitably split up contiguous sections of his original 

Some questions of little importance concern the relations of the pro- 
totypes C, Aa, and Ab. In "The Scott. Geogr. Mag.", vol. XXX, p. 70, 
we have suggested the equation Amisia C — AHson Aa. In the same 
volume, p. 621, we have withdrawn this suggestion. Ptolemy's town 
Amisia near the homonymous river actually existed and still exists as 
Ems on a homonymous river in Hessen- Nassau; it belongs to Prot. Ab. 
The fortress of Luppia, assigned by us to Prot. C, ibd. p. 70, may pos- 
sibly also belong to Ab. 

A more important alteration of our views affects the demarcation of 
Prot. C against Prot. D. On the Ptolemaic map, Prot. C is most ob- 
viously betrayed by its constant eastward displacement. The same dis- 
placement characterised parts of our assumed Prot. D, whereas others, 
such as the "Swabian" tribes of Angles and Langobards, are displaced 
in exactly the opposite direction. 

The parallel displacement would of course make the distinction of the 
two prototypes difficult, but we believed that we had found firm ground 
in the duplicate Chaimai = Kamauoi. As the tribe Chaimai stands 
among . Belgian towns evidently belonging to C, we assigned it to this 
prototype. Consequently, its alter-ego Kamauoi would belong to D^ and 
this prototype would hence lay claim to the entire surrounding milieu of 
tribes: Chairusikoi, Chattai, Tubantoi, etc. 

Having divided the prototypes in this way, we further searched for 
physical lines which might conceal the original framework of C and D. 


And we suggested that the Ptolemaic mountain Melibokos might be 
regarded as representing the original Rhine of both prototypes, only in 
the opposite direction : in C, east- west must be reconstructed as north- 
south^ whereas in D -it would be south-north. See figures 6 and 7 in 
the first article, vol. XXX, p. 57. 

The above theory of division would be certain, if it could be taken 
for granted that the Ptolemaic Chaimai belonged to Prot. C. Later, how- 
ever, we noticed that their pertinence to C is less certain than we had 
thought at first: in spite of their eastward displacement, they might also 
belong to Prot. Aa. The Ptolemaic North-Sea tribes derived from Prot. 
Aa may not necessarily all be correctly localised. As a matter of fact, 
Ptolemy places the Parisians too far south, practically at the place of 
the Chamavi, and so it is possible that they have displaced the 
latter towards the east, no matter whether this displacement occurred 
already in the Prototype Aa, or whether it was due to the Ptol. con- 

As soon as we assign the Chaimai to Prot. Aa, nothing prevents us 
from regarding Prot. C as owner of the duplicate Kamauoi, and of the 
entire surrounding milieu. Prot. D, on the other hand, would lose most 
of its contents, being reduced to the trinity of Swabians, i. e. Semnones, 
Aggeiloi, Laggobardoi. 

Considering the two alternatives, we feel obliged to decide in favour 
of Prot. C, declaring this prototype owner of almost all the displaced 
Ptolemaic tribes between the Rhine and the Elbe, except the Swabians. 
Our principal reason is the fact that the correspondence with the Tabula 
Peutingeriana will only become complete, if we may attribute to C the 
Kamauoi-Chairusikoi = the Chamavi-Chrepstini Tab. Peuting. 

d. Topographic Correctness. 

As we shall see under the heading "Literary milieu", Prot. C was a 
so-called itinerary, i. e. a road-map showing the distances between a 
series of towns. Such maps, like our modern schematic representations 
of railway systems, do not pretend to offer a correct topography. The 
Tab. Peuting. is a classical exemple of the prevailing distortions. Corre- 
spondingly, there are several traces of bad topography in Prot. C. 

The tribes are distributed in a confused manner, as in the Tab. Peuting. 
(i) Abrinkatuoi, (3) Namnetai, (2) Redones, instead of i, 2, 3; Karitnoi 
south of Vangiones, *Chattvaroi south of Kasvaroi, Kamauoi beside 
Chairusikoi and Chattai, cf. the Chamavi beside the Chrepstini on the 
Tab. Peuting. The distance between Nouaision and Bogadion (Bagacum) 
is shortened. On the other hand, the distance between Mediolanion and 


Leufana is largely exaggerated. *Bagacum ought to lie south-west of 
Nouaision, not north-west. 

It must, however, be remembered that the Ptol. constructor may 
have deteriorated the map, as he seems to have done by introducing 
the broad chasm between the sections north western Gaul and Germania 

In spite of the confusion, parts of the map seem to have been not 
so bad after all. The line *Vangiones, *Arlaunon, *Tungri (Tenkeroi), 
Nouaision, Teuderion corresponds fairly well to the actual positions oi 
Worms (capital of the Vangiones), Arlon, Tongern, Neuss, Tiiddern. It 

seems that Prot. C had not yet assumed the 
extremely oblong shape which deforms the 
Tab. Feuting. The Rhine was probably 
represented on the basis of observation of 

Leufana o 


o Nouaision ... ^ i. ^i, i- 

its various curves, and not as a smooth line, 
as was the case on the Tab. Peuting. and 

*Usipoi in Prot. A. A zigzag line is implied by the 

localisations of the fixed points indicated in 

the accompanying diagram. It corresponds to two actual curves of the 

river, the one between Strassburg and Bingen, the other between Neuss 

and Nimwegen. 

e. Statistical Features. ^ 

Prot. C, like the Tab. Peuting., contains mainly towns, but also several 
tribes. In the invaded north-western German section of Prot. Aa, Prot. C 
thus ''supplies a want", as these regions were in reality almost completely 
bare of towns. 

On the other hand, Prot. C enriches south-western Germany with a 
series of tribal names, whereas the due local prototype Ab contains no 
samples of this category. The present selection of names in Prot. C 
seems somewhat accidental or arbitrary, but this fact may to a great 
extent be due to the Ptol. constructor. It is worth noticing that all 
traceable towns of Prot. C in the western section possess a certain im- 
portance, three being tribal capitals, and the fourth a flourishing mercan- 
tile centre, the present Rouen. On the Tab. Peuting. two of these are 
distinguished by towers, viz. Rouen and Rennes. 

As to the names of tribes, the selection may have been somewhat 
arbitrary from the very beginning. We shall see later on that it is 
reproduced almost unaltered by the Tab. Peuting. But Prot. C at least 
in some points is more complete than both the Tab. Peuting. and the 
Itinerarium Antonini. The * Vangiones, Abrinkatuoi, Redones, Morinoi 
of C are missed in both of the latter documents. And, if we are right 
in assigning to C the Kamauoi and their surroundings, the prototype 


would have contained a fairly copious representation of tribes in western 
Germany, whereas the selection of the Tab. Peuting. is more fragmentary, 
and no German tribes occur in the parts concerned of the Itin. Anton. 

f. Occurrence of Duplicates. 

The distribution of duplicates has been somewhat altered, owing to 
the modification of our theory. We now assume the following series: 
Namnetai C = Samnitai A^ Ratomagos C and A, roMorinoi C = Morinoi A 
(not in all MS. atlases), Vargiones C — Vaggiones A, Askalingion C = 
Askiburgion A, Kamauoi C = Chaimai A. The two Marionis can no 
longer be regarded as authentic duplicates, as the one belonging to C is 
rather a distortion of Matilone on the Tab. Peuting. We have withdrawn 
the identification of Intouergoi C and Triberoi A (Trevagroi Strabo), 
preferring the combination with the Nitiobroges of the Tab. Peuting. Also 
the equation Feugaron C = Tungroi A seems too questionable. 

g. Linguistic Marks. 

In "The Scottish Geographical Magazine", vol. XXX, p. 71, we pointed 
out Latin residuals in Prot. C, such as the nk, ng in Kondeoui^^on, I;2^ena, 
Abri;2y^atuoi, Askaliw^ion. It may be added that Leufan^a: points towards 
the vulgar Latin form Leuefan<7, Tab. Peuting. ; a Greek MS. would 
scarcely have dropped the final nasal as early as the second century A. D. 
Our new conjecture Albanianis, Tab. Peuting. = *Albis amnis, Prot. C, 
suggests that the prototype would have been read and interpreted in 
Latin. Also the equation F-abiranon = Foro Adriani, Tab. Peuting., seems 
to point towards Latin types. 

A pre-Ptolemaic trace of Greek editorial language is perhaps the 
erroneous spelling XA/T0Yi2P0I < XATTOYAPOI; the Latin corre- 
spondence AE = Greek AI would not so easily be derived from TT. 
The CO in Vargiones, Inkriones may equally originate from a Greek pre- 
liminatory stage; otherwise, Ptolemy constantly writes -ones in Gallic and 
Belgian names, except in Keutrones that is placed within Italian territory. 

h. Literary Milieu. 

Prot. C is most closely related with the Tab. Peuting., but has also 
special affinities with the Itin. Anton. 

I. Common Affinities. 

Towns: Kondate, Ratomagos, *Bagakon, Koinoenon (Garuone), 


2. Affinities with the Tabula Peutingeriana. 

Tribes: Namnetai, Intouergoi, Vargiones:?, Karitnoi, Brukteroi?, 

Kamauoi, Chairusikoi (Kasvaroi?). 
River: Amisias (for Anatius). 
Towns: Askiburgion (for cAspingium), Nabalia, Fleum, F-abiranon, 

Matilone, Albis, Leufana, Askalingion. 

3. Affinities with the Itinerarium Antonini. 

Tribal district: Germania. Tribe: *Tenkeroi (=: Tongri). 
Towns: Mediolanion, Teuderion, *Arlaunon. 

The correspondence regarding the names Leufana, *Arlaunon, Teu- 
derion, Mediolanion is worth noticing, because the classical records of 
these four names are limited to the three authorities mentioned. 

We have mentioned above that two of four Gallic towns in C have 
vignettes with towers on the Tab. Feuting. — a circumstance which points 
towards statistical parallelism. We have likewise mentioned the close corre- 
spondence between the tribal names of Prot. C and those of Tab. Feuting. 

Towards the east, both descriptions extend as far as to the Cherusci 
and no further. In the part of Gaul situated north of the Loire, the 
Tabula contains hardly any additions to the stock of Frot. C. We notice 
only Veneti, Osismi, Franci; the last-mentioned name must be regarded 
as added after Ftolemy's times, as it existed scarcely before our era and 
occurs never in literature before the publication of the Tabula. 

If Fabiranon is correctly interpreted as Foro Adriani, Prot. C would 
originate from the times of the Emperor Hadrianus, i. e. after 117, or at 
least its last edition would belong to this period. 

i. Examination of Details. 

On practical reasons, the details concerned have been discussed under 
the heading "Ptolemaic localisation". 

Artaunon confirms the present forms of the name, French Arlon 
(occurring since 870, according to "La grande Encyclopedie"), Flemish 
Aarlen. The form Orolauno of the Itinerarium Antonini is of similar age, 
appearing both in inscriptions and documents. Perhaps the ambiguous 
spelling *Arlaunon & Orolauno denotes an old contrast between Gothonic 
and Celtic pronounciation, as in Masa versus Mosa (the Meuse), Wasgen- 
wald versus les Vosges, etc. 

j. Conclusion, 

The main interest of Prot. C is merely literary, consisting in the fact 
that it 4ielps to illustrate the genesis of the Tab. Feuting. and the Itin. 



a. Summary of Contents. 

Prot. D is only traceable as a fragment. It is a local description of 
the Swabian group, containing only tribes. A duplicate name occurs in 
Aa (or perhaps two). There are Greek marks. Affinity with Strabo and 
Tacitus. Cf. Fig. 2. 

b. Ptolemaic Localisation. 

The Ptolemaic constructor has introduced Prot. D into the interior 
part of the Germanic territory belonging to Prot. A. From the physical 
point of view, D has left no trace, but its presence is apparent from the 
large displacement of well-known names. The Semnones are fairly speaking 
correctly localised, but the Angles have emigrated from the Baltic shore 
to Thuringia, and the Langobards from the' Elbe to the Rhine. Prot. D 
perhaps also contained the Ptolemaic Farodinoi in Mecklenburg = the 
Charudes from northern Jutland. — The position of the Swabians about 
the middle Rhine may be an inheritage ("apochronism") from the year 
58 B. C, derived from Caesar who describes a Swabian attack against 
this region. The Angles seem to have obtained their place in interior 
Germany owing to erroneous identification with the *Angrivarii of Prot. C, 
cf. p. 119. 

c. Definiton of Limits. 

The addition of "Sveboi" is the main characteristic of B. Incorrect 
arrangement distinguishes D from the elements of Aa in north-western 
Germany. Ptol. displacement from east to west distinguishes D from the 
elements of C with the Ptol. displacement from west to east. For further 
particulars cf. § 24. 

d. General Topographic Scheme. 

Not traceable. Prot. D seems to have been a descriptive text, 
no map. 

e. Statistical Features. 

Only tribes are traceable. The Angles are emphasized (as sole repre- 
sentatives of the Nerthus group, cf. under h.). 

f. Occurrence of Duplicates. 

Laggobardoi = Lakkobardoi Aa. Perhaps further Farodinoi = 
Charudes Aa. 

126 Ptolemy's maps of northern Europe 

g. Linguistic Marks. 

Latin marks are not traceable. 

The combination n£- is written in corrrect Greek manner as ^-g- : 
La^obardoi, A^^eiloi. Cf. the contrasts in the surrounding prototypes: 
LAiTiTOBARDOI Aa (pointing to an original "^LANCOBARDl), Kngn- 
varioi Aa^ Askaliw^ion C^ Asaw^a, Si;/^one Bi^ Marvi/^^oi B2^ regular 
ng in Acde, 

Instead of Semnones we might expect the spelling Semnones, as used 
by Strabo. But even the Senonic Gauls in Italy are by Ptolemy written 
Semnones, and Dio Cassius has the same spelling, LXVII, 5. Evidently, 
the Semnones as an important tribe had a relatively fixed orthography, 
which preferred the 0, because the Greeks knew the name through the 
medium of Latin. 

h. Literary Milieu. 

Prot. D recalls Strabo and Tacitus, the only two authors who emphasize 
the Swabian group in a similar manner. The designation of the tribes 
round the Elbe as Swabians must be referred to the establishment of 
King Marbod's great Swabian Empire about the beginning of our era. 
The Semnones and Langobards are directly mentioned as Marbod's sub- 
jects or allies, cf Tacitus, "Annals" II, 45 (17 A. D.). Also the Angles 
as neighbours of the Langobards may have belonged to Marbod's vassals. 
— As the Angles were no Swabians in the ethnic sense, the continued 
designation ''Sveboi Aggeiloi" must be regarded as a "political apo- 
chronism". This antiquated designation, together with the solitary in- 
stance of the name of the Angles, constitutes a typical affinity with 
Tacitus. The antiquated "Swabian nationality" re-appears in his description 
of the Aestui, who were in reality no Gothonic nation, but belonged to 
the Lithu-Prussian group; cf. Strabo VII, 290, who represents Marbod as 
king of the "Sibinoi" i. e. the Sudines in Prussia. 

The selection of tribes also betrays a marked affinity between Prot. D 
and Tacitus. The Farodinoi D (?, Charudes Aa) may be re- discovered 
in the Tacitean Suardones or Suarines who belong to the Anglian group. 
In Prot. D the Swabians are represented by the Semnones, the Lango- 
bards, and the Angles. In the "Germania" of Tacitus, the Semnones 
and Langobards are named first, and emphasized as the most prominent 
representatives of the group. The Angles belong to a special group of 
Swabian tribes, worshipping Nerthus, and mentioned directly after the 
Langobards. It is true, the Angles are not given by Tacitus special 
prominence over the other six Nerthus-peoples, but we do not require 
the evidence of the Roman author to realize that they were in reality 
the leaders of the community. We may say that the combined evidence 


of Prot. jD and Tacitus points towards a source that valued the Angles 
according to their actual prominence which remained otherwise concealed 
in historical literature till the times of Procopius, 6th century. 

i; j. Examination of Details; Conclusion. 

In spite of all Ptolemaic confusion, Prot. D contains one highly 
valuable detail, viz. the name of the Angles. We are informed that they 
are the neighbours of the Langobards towards the north or north-east, 
— - a statement which is made nowhere else in classical literature. Of 
course, we must remove the Langobards of D back to the place of the cor- 
rectly situated alter-ego in Prot. Aa, the "Lakkobardoi" in the present 
Bardengau round Liineburg. Consequently, the Angles must be placed 
north or north-east of the region, i. e. fairly in their traditional home- 
stead, the district of Angel in Slesvig or South Jutland. Thus Prot. D, 
far from contradicting the venerable Bede, in reality proves his most 
valuable supporter. The unanimous evidence of local nomenclature, 
linguistic features, reHgious institutions, and genuine English, Danish and 
German tradition, is thus crowned by the hithertho missing element, the 
evidence of classical cartography. It is needless to discuss the matter 
any more^). 


a. Summary of Contents. 

Prot. E 8l F are collective maps, describing eastern Germany, Sar- 
matia Europaea, Sarmatia Asiatica, and Scythia, containing all sorts of 
geographic categories; F is besides marked by a system of "ethno-topic 
denomination". The prototypes are duplicates of each-other; scattered 
duplicates occur in Aa, Acde, Bt, B2, E has Latin marks (Sarmatai 
instead of Skythai F), but seems fo have been translated into Greek be- 

*) We must here urge Chadwick's warning against rejecting the well verified native 
tradition in favour of the somewhat older, but peripheral evidence of an inaccurate classical 
geographer like Ptolemy, As long as the genesis of Ptolemy's work remained unexplored, 
his evidence in the Anglian question was practically worth nothing. — We may add one 
hitherto ignored piece of traditional evidence concerning the Angles. The Quedlinburg An- 
nals, written in the nth century, say ad annum 445: "The Angles, conducted by their king 
Angling, leave the country of the Danes". 

128 Ptolemy's maps of northern Europe 

fore the stage of Ptolemy. F has only Greek marks. — E & F are 
executed after the introduction of a well established Roman amber trade 
with the Baltic regions under the reign of Nero. Affinity with Pliny, 
including antiquated Herodotian names. Cf. Figures 3, 17, 18, 19, 30, 31. 

b. Ptolemaic Localisation. 

E is totally displaced, F is correctly localised. 

The Ptol. constructor has compressed Prot. E within the sections 
called Sarmatia Europaea and Asiatica, partially owing to the fact, that 
the Scythians were in this prototype called Sarmatai. E has been turned 
round, so that west becomes south, and east becomes north. Thus the 
Germanic part occupies the south-western edge of Sarmatia Europaea, 
whereas the remainder of the prototype forms the most northerly peri- 
phery of the Sarmatian sections. 

Through this displacement^ the eastern Baltic coast was enriched with 
some three or tour rivers, originally flowing into the Black Sea, viz. 
Rhudon = Rhode, Turuntes = Karkinites (?), Chesinos = Acesinus. The 
fourth river, Chronos, may also be a transplanted one, or it may be a 
really Baltic river, originating from Prot. F. The river pAsiakes £ = 
Axiakes F still keeps its place in the region of the Black Sea (together 
with the towns Leinon, Erkabon, and Trabana = Leianon, Sarbakon, and 
Tabana FY). 

The displacement was to a great extent' due to the misinterpretation 
that the Baltic coast was taken for the river Vistula. This fact appears 
clearly from the Ptolemaic tribes, localised east of the Vistula: Ombrones 
— Ambrones, the campanions of the Cimbri and Teutones; Auarinoi = 
the Varines, a well-known tribe from Mecklenburg; Frugundiones = 
Burgundians, inhabitants of Pomerania. 

Prot. F meets Prot. Aa in the Baltic region, cf. the duplicates 
Teutones- Auarpoi F = Teuton . . Ouirunoi Aa. Correspondingly, F meets 
Ac near the Black Sea, cf. Harpioi with town Harpis F = Karpianoi Acde. 
In the Baltic region, the details of F are distributed among those of Bi 
& B2 so that they are not easy to discern. 

We suppose that Prot. F has been enriched with the contents of Sk, 
i. e. Scandia, before both of these prototypes were amalgamated with the 
collective prototype A. Only through this assumption, we are able to 
explain the occurrence of the name Finnoi in Prot, E. As E appears 
generally as an extract of F, the description, of Scandia with the name 
Finnoi seems to have been incorporated with F, before the extract was 

^) Most of the identifications are suggested by C. Mtiller. 


d. General Topographic Scheme. 

Both E and F contained coasts of the Black Sea and of the Baltic. 
The design of £ seems to have been so indistinct that the coast of the 
Black Sea might be mistaken for a mountain-chain in F, — The latter 
prototype was an excellent map and may be regarded as the main 
foundation of the Ptol. maps of Sarmatia and northern Scythia. Here 
we notice, as Miillenhoff remarks, the Caspian Sea for the first time cor- 
rectly represented as an inland water and not as an inlet of the northern 

The design of mountains in F seems to contain true observations of 
the low ranges of hills running through eastern Europe: Peuke = Lysa 
Gora in Poland, Wendian Mountains = the hills of Suwalki east of Prussia, 
Bodinian-Alanian-Ripaeean Mountains = western Russian Range, Hyper- 
borean Mountains = Waldai Hills. However Sadowski maintains that the 
so-called mountains are simply theoretical expressions of water-sheds, 
s. "Die Handelsstrassen der Griechen und Romer durch das Flussgebiet 
der Oder-Weichsel". 

Finnoi = Finns in Finland or Scandinavia. — The coast of the Black 
Sea in F seems to have been mistaken for the mountains of interior 
Sarmatia F, whereas these same mountains, as they appeared in £^ were 
in return mistaken for the Baltic coast in F. Thus a complete turning 
upside down was effected. 

The decoration of the utmost north of Europe with numerous anti- 
quated or fabulous Herodotian tribes, such as Melanchlainoi and "Horse- 
foot-men", seems to be a sort of intentional swindle, committed in order 
to conceal the Ptol. constructor's ignorance about this extremity of the 

The Ptol. constructor has treated Prot. F quite otherwise than its 
alter-ego E. He localised F correctly, and he could hardly avoid it, 
owing to its evidently distinct and finished design. It has been amal- 
gamated with Prot. A without any trace of inconsistency. And probably, 
we owe to Prot. F a great deal of the physical framework in the eastern 
parts of Ptolemy's atlas. As might be expected, Prot. F has not com- 
pletely escaped deterioriation through the Ptol. constructor, — one such 
case will be mentioned under e — ; still such cases are of minor im- 

c. Definition of Limits. 

Taking it as a whole, £ and F are easily distinguished from each-other, 
partially through the series of duplicates, partially through the contrast of 
wrong and correct localisation. An additional criterium is the designation 
"Sarmatai" in £, replacing "Skythai" in F; further the system of "ethno- 



topic denomination" of F, cf. under e. The occurrence of the denomin- 
ation ^'Sarmatai" in E is connected with the fact that the Ptol. con- 
structor has limited this prototype to the so-called Sarmatian sections of 
the atlas, not only in Asia (cf. above p. 128), but also in Europe. E ge- 
nerally occupies the most northerly periphery which was left blank in F. 
Owing to this circumstance, the confusion of E and F is comparatively 
little. However, in western Sarmatia there is a somewhat large area of 
confusion. The displaced Baltic tribes of ^ — Ombrones, *Ouarinoi, 
Frugundiones, Sulones, Finnoi — stand south of their correctly localised 
alter-egoes in F. Likewise, the names pAsiakes, Leinon, Erkabon, Tra- 
bana of E, belonging originally to the regions near the Black Sea, are 
placed in the middle of elements originating from F. 

The displaced Baltic detachment from E stands in an isolated posi- 
tion, in sharp contrast to the correctly localised names on both sides: 
those of Prot. Bi in the west, and those of Prot. Acde in the east. The 
tribes Ratakensioi and Kotensioi inside the Dacian area of Acde may 
originate from E. Otherwise, Prot. E collides with no prototypes except 
its own alter-ego F» 

e. Statistical Features. 

Prot. E has a less copious selection of details than Prot. F. The 
complete absence of towns in the northern parts of E contrasts with the 
copious lists of towns in the neighbouring Prot. Bi, and also in the 
Ptol. description of Jazygia. 

Prot. F, as we have repeatedly mentioned, is marked by the system 
of "ethno-topic denomination". Its western vanguards are: the Venedai 
with Venedian mountain and gulf, i. e. represented as inhabitants of the 
eastern Baltic coast ; the Peukinoi with the mountain Peuke ; the southern 
outpost of the Peukinoi on the island of Peuke in the Danubian Delta; 
the Harpioi with the town Harpis. 

The presence of "ethno-topic denomination" at a Pre- Ptolemaic stage 
appears from the following correspondences, noticed by C. Miiller: 

A. Caucasian Region, B. Siboian Region. 

la. Paniardis, district lb. Paniardoi, tribe 

2a. Konapsenoi, tribe 2b. Konadipsas (Kanodipsas), district 

3a. Korax, mountain. 3b. Koraxoi, tribe. 

The two lists of names originally must have formed a chain of "ethno- 
topic denomination", but in Ptolemy's work they have been split up, 
list B being transplanted far away from its proper place, and hence it 


appears that Ptolemy did not invent the system of "ethno-topic deno- 
mination", but found it ready-made in an earlier work. 

In the neighbouring prototypes, the cases of "ethno-topic denomin- 
ation" are so rare that they may be regarded as accidental. We notice 
e. g. within the area of Prot. Aa these three cases: Kimbroi with Kim- 
brike Chersonesos, Saxones with Saxon islands, Virunoi with town Virunon. 

f. Occurrence of Duplicates. 

The duplicate series of Prot. E and F (Fig. 24) is very long, containing 
some 24 pairs of names. It could scarcely be expected that parallel 
chains of such a length would agree completely in the order of links. 
Yet the approximate agreement of the series — especially in the upper 
lists (Auarinoi .... Hippofagoi Sarmatai £ = Auarpoi .... Hippofagoi 
Skythai F) — must be called surprising and excludes any chance of 
accidental coincidence. Since our first article in *'The Scottish Geogra- 
phical Magazine" we have suggested a new equation: Gelonoi E = 
Geiounoi F (Cod. Palat. 191, instead of the hitherto accepted reading 
Geouenoi). There are also some duplicates or triplicates which serve as 
means of distinguishing E and F from the other prototypes. 

Auarinoi E, Auarpoi F = Ouirunoi Aa. 

Teutones F = Teuton(-oaroi) Aa. 

Harpioi with town Harpis F = Karpianoi Acd^. 
Ratakensioi E} = Rakatriai Bi, Rakatai B2. 

Kotensioi (Kontekoi) E} = *Kotnoi Bi, *Koteinoi B2. 

g. Linguistic Marks. 

The final editorial language of both E and F seems to have been 
Greek. Cf. the following peculiarities: 

Spelling au instead of the Latinising aou: S<a:^aroi E = 'Nauaroi F. 
Misreading au for the Greek ou: ^^^arinoi E = ^^^arpoi A, contrasting 

6^«irunoi Aa (<Z Viruni). 
Misreading Pasiskes E (=■ Axiakes F) = *Potamos ^siakes. 
Constant "omega" in the termination -ones: Ombr^nes, Frugundi^nes, 

Sul^nes, Karb^nes, Vibi^nes, Gel^nes, Igylli^nes E, Gyth^nes, Kari^nes, 

Eluai^nes F. 

But in E we notice traces of a Latin pre-existence. The most con- 
spicuous is the term "Sarmatai" instead of the Greek synonym "Skythai", 
see above. E. g. the Herodotian Basilikoi Skythai appear as Basilikoi 
Sarmatai, etc. If this translation is omitted in the name Exobygitai = 
Hamaxobioi Skythai F, it seems due to the circumstance that the name 



had in E become unreadable at an early stage. The distortion itself 
seems to point towards a Latin document: the misreading -BY- would 
originate from a Lation -BII rather than from the Greek -BIOI, and 
-GITAI from a vulgar Latin *SCITHAE rather from the Greek SKYOAL 
— The name Portakra in the Crimea, probably originating from E^ con- 
tains the Latin word portus, "harbour". 

We have mentioned in § 16 that the Latin traces of Prot. A form a 
marked contrast to the Greek traces in the duplicates on the western 
frontier of Prot. F\ Latin correcture *"Vari" over *"Viruni" Aa facing 
the Greek misreading Auarpoi = Ouarinoi F\ Latin termination in Kar- 
pianoi Acde facing the ethno-topic couple Harpioi-Harpis F. 

h. Literary Milieu. 

The entire literary milieu of Prot. E and F cannot be investigated 
here, as it would lead us too far into the history of Asiatic geography. 
It must be sufficient to state the conspicuous affinity with the sphere of 

This affinity appears perhaps most strikingly at the western edge of 
the area concerned: Auarinoi-Frugundiones-Sulones E = Auarpoi-Bur- 
guntes-Gythones F seem to be identical with Pliny's list of "Vandilian" 
tribes, IV, 99: Burgundiones-Varinne-Gutones. It is true that Miillenhoff 
in his "Germania antiqua", p. 93, eliminates "Varinne" as a distorted 
duplicate of the immediately following Charini, the Harii of Tacitus. We, 
however, cannot admit his opinion as justified; for "Varinne" is not far 
from the well-known tribe-name Varini, and the existence of a traditional 
Baltic list, Varini, Burgundiones, Gutones, seems confirmed through the 
coinciding evidence of three authorities, viz. Prot. E, Prot. F, and Pliny. 

In the description of the Maeotian coasts, the affinity between the two 
prototypes and Pliny is equally conspicuous. 

1. Common affinities. 

Tribes: Neuroe, Hamaxobii, Rhoxolani, Aorsi, Geloni. 
Rivers: Axiaces, Pacyris, (Carcinites = Turuntes E). 

2. Pliny and Prot. E. 
Tribes: Basilidae, Agathyrsi. 

Rivers: Rhode, Acesinus (C. Miiller, I, 412). 

3. Pliny and Prot. F. 

Tribes: Budini, Tyragetae; colony of Cares = Karoia (C. Miiller, 

1, 418). 

Towns: Nauarum, Carcine, Taphrus. 

Rivers or Gulfs: Buces, Gerrhus, Hypanis, Panticapes, Coretus = 
Poritos, sinus sAggarius = Agaros. 



The geographical work, from which Pliny extracts his description of 
the Maeotian coasts, is subjected to detailed examination by Miillenhoff 
in his '^Deutsche Altertumskunde", III, 53 seq. Mela used the same 
work. It is marked by the presence of numerous Herodotian names 
which were in Mela's and Pliny's times already antiquated. We re-discover 
most of them in Prot. £, whereas an editor of Prot. F has evidently 
tried to reduce the anachronistic character by eliminating antiquated 
names, apart from some residuals such as Bodinoi and Geiunoi = 
Gelonoi E. 

Sometimes we notice that Pliny and the Ptolemaic prototypes represent 
the same development leading away from the original source. E. g., all 
of the three authorities add new names, such as Hamaxobii, Rhoxolani, 
Aorsi. The Agathyrsoi are by £ placed among the Maeotian tribes, 
corresponding to Mela and Pliny, whereas Herodotus placed them in 
Dacia. The Herodotian name Hypakyris is unanimously written without 
the initial syllable Hy-. The Neuroi appear with a town Nauarum Pliny 
= Nauaron F] the same new vocalisation appears in the *Nauaroi 
(Sauaroi) of F. 

This line of development seems to have been continued^ by F and F, 
introducing several times a contrast to the stage of Pliny. E. g., the 
Herodotian river Hypakyris is still by Pliny preserved as the river 
Pacyris, whereas F and F turn it into a race-name : the tribe Pagyritai F, 
= the town Pasyris F (C. Miiller, I, 432). Whereas Pliny leaves the 
Herodotian Neuroe unaltered (beside the town Nauarum with the new 
vocalisation), F writes not only Nauaron, but also Nauaroi = Sauaroi E. 

Whereas the affinity between F, F, and Pliny appears at the first 
glance, there are generally no traces of special affinity between the two 
prototypes and Tacitus. We miss almost entirely the tribes, mentioned 
by Tacitus as inhabitants of north-eastern Europe: Aestui, etymologically 
=: Esthonians, with ''lingua Brittannicae propior" (probably a disguised 
notice of the Pruteni or Prussians); Sitones, governed by queens, i. e. a 
disguised notice of the Quaenes; Hellusii; and Etiones, i. e. the Jptnar of 
Norse tradition. . 

However, Prot. F contains at least one marked affinity with Tacitus, 
namely the presence of the Finns, who are not mentioned in those books 
of Pliny which have been preserved. 

i. Examination of Details. 

In spite of all confusion, Prot. F contains at least one valuable topo- 
graphic detail, viz. the name of the Ombrones. This tribe is mentioned 
nowhere else in geographical literature, but we recognize it as identical 
with the historical Ambrones, the companions of the still more famous 


Teutones and Cimbri. Cf. Miiller's edition, I, p. 424. Through E we 
are infornned about their localisation. They appear south of the Auarinoi, 
read : west of the Ouarinoi in the present Mecklenburg, — a tribe which 
is known among the Anglian tribes worshipping Nerthus. — We may- 
identify the Ambrones with the present Amrings, living on an island 
west of Slesvig called Amrum, in mediaeval times Ambrum. Perhaps the 
name has also some connection with Imbrae, as the island of Fehmern 
was called in Old Danish. In the Old English epical catalogue Widsith 
the tribe -re-appears as Ymbre, and Welsh authors such as Nennius still 
used Ambrones as synonymous with Saxons. 

j. Conclusion. 

The prototypes E and F must be called well verified both from topo- 
graphic, linguistic, and literary points of view. 

Prot. E contains only one valuable individual element, viz. the tribe- 
name Ombrones, localised *west of the *Ouarinoi. Otherwise, its value 
consits in the thoroughgoing confirmation which it affords to the duplicate 
prototype F. 

The latter, on the other hand, is one of Ptolemy's most valuable 
sources. We notice especially the correct representation of the Caspian 
Sea as an inland water. 



a. Summary of Contents. 

Prot. Sk is a special map or description of the Scandinavian Peninsula, 
containing tribes only. — No duplicates, except Finnoi in E. — Greek 
marks. A limited affinity with Tacitus. Cf. Figures 3, 20. 

b.; c. Ptolemaic Localisation; Definition of Limits. 

It seems that Prot. Sk was amalgamated with Prot. F before the 
Ptolemaic stage, cf. § 24 b. The Ptol. constructor introduced Sk into 
the outlines of the Peninsula of Scania, as represented in Prot. A (from 
the local prototype Aa). The area of Scania was of course far too narrow 
to contain the seven Scandinavian tribes of Sk, and therefore most MS. 
copies of the Ptolemaic atlas simply leave the map blank. This is one 
of our principal reasons for concluding that Sk must have a different 
origin from the Scandian coast design of the atlas. Another reason will 
be found in the commentary on the literary criteria. 

§ 27- LOCAL PROTOTYPE 5;fe 135 

d. General Topographic Scheme. 

The physical nature of the Scandinavian Peninsula makes it self- 
evident that this country must have been described in a separate proto- 

e. Statistical Features. 

Prot. Sk contains only tribes. These are well selected as they re- 
present generally the more important inhabitants of the peninsula. The 
*Finaithoi, or people of F'inveden, would perhaps not seem important 
from a modern point of view, but as a matter of fact they appear again 
in the next detailed description of Scandinavia, namely that which is 
given by Jordanis in the 6th century; we may identify them with the 
primaeval inhabitants of the entire province of Smaland. It is highly 
remarkable that the Norwegians are represented by the inhabitants of 
Hedemarken: for this province is actually the most fertile in the whole 
of Norway and must have been an original centre of Gothonic race 
within this country. 

f. Occurrence of Duplicates. 
Finnoi, re-appearing in E, cf. § 26. 

g. Linguistic Marks. 

All marks of Prot. Sk are Greek. 

Spelling eu^ not the Latinising eow. l^euonoi. 
Misreading ou for au: Goutsx. 
Misreading au for ou\ ¥auondi\ = *Souionai. 

Spelling -ones with "omega", not with "omikron": Leu^noi, Dauki<?nes. 
Thus the original document seems to have been Greek from the very 
beginning, never subjected to Latin transcription. 

h. Literary Milieu. 

The contents of Prot. Sk are quite unique. Only few or vague 
affinities are found in classical literature. 

Already Mela knew the '^island of Codanovia", i. e. Scadinavia, Scan- 
dinavia = Scandia. But he seems to have known little more than the 
bare name. 

Prot. Aa has a relatively exact description of the coast of Scania, 
but nothing else, cf. § 20. 

Pliny has an essentially wider knowledge about the Peninsula, evidently 
dating from the lively mercantile intercourse with the Baltic amber coast, 
established under the Emperor Nero. Not only does Pliny repeat the 

136 Ptolemy's maps of northern Europe 

names Scandia and Scadinavia, already known by Aa and *Mela, but he 
supplies several new details. He also seems to have known a description 
which represented the "island" of Scandinavia correctly as a peninsula, 
— only he did not recognize the identity of Scadinavia with the penin- 
sular country mentioned. The peninsula, he says, contains the immense 
mountain Saevo = Kolen and Dovre in Norway, IV, 99. The dimen- 
sions of the "island of Scandinavia" are much better known by Pliny 
than by Ptolemy: it is not that tiny bit which appears on the map, but 
a country which rivals the remainder of Europe in size. This is re- 
presented as the opinion of its inhabitants, who only cover a portion of 
the island, although they embrace 500 counties (pagi). The name of the 
inhabitants is mentioned: "Hilleuionum gente", perhaps to be amended 
into "ilia Suionum gente", as there follows a relative sentence^). Besides, 
Pliny reports fabulous stories about the Scandinavian fauna. 

Pliny's correct ideas of the dimensions of Scandinavia re-appear in 
Prot. Sk. And, as we have pointed out, the name of the Swedes is 
perhaps common to the two authorities. But otherwise, the milieu is 
rather that of Tacitus and his age. 

Tacitus is strikingly well informed about the Scandinavian Peninsula. 
His detailed data seem especially remarkable, when compared with his 
vague ideas about the Cimbric Chersonese. Evidently, the wide exten- 
sion of the Tacitean horizon over the Scandinavian Peninsula is due to 
the continued and growing intercourse of the Romans with the Baltic 
amber coast ^). In Scandinavia, Tacitus mentions only two nations, 
Swedes and Sitones; besides, his Finns may be assigned to the same 
sphere. Probably, he knew more, but did not regard it as adviseable to 
fill his brief survey with mere names. In return, the nations mentioned 
are relatively exactly described. Tacitus records the Swedish kingdom, 
the Swedish navy, and a series of customs which evidently point towards 
the national Swedish cult of the male Nerthus or Freyr. The Sitones, 
according to Tacitus, are governed by queens. It is a popular tale, 
originating from their Scandinavian name, Kvaener. In mediaeval literature, 
the country of the Kvaener was called "terra feminarum", i. e. "women- 
land". The Kvaener are in reality Finns, although Tacitus regards them 
as Swabians, i. e. as a Gothonic nation. Finally, Tacitus describes the 
Finns, whom he seems to have regarded as living on the continental 
coast opposite Scandinavia. Their poor living and savage customs are 
described in a detailed way. 


^) Suggested by us in Salmonsen's "Illustr. Konversationslexikon". Also suggested 
by LafBer. 

^) Cf. the word lukarna-staki ("candle-stick") on the island of Gothland, borrowed from 
the Gothic lukarna-staj)a = Latin lucerna. It is a most striking evidence of the influence 
of Roman trade on Scandinavia. 

§ 2 7- LOCAL PROTOTYPE Sk 137 

If we compare Prot. Sk with Tacitus, we find both a general and 
special accordance. Both authorities have more exact ethnic details 
from Scandinavia, than Mela og even Pliny had. Both authorities know 
of Swedes and Finns. The latter nation is unknown apart from Tacitus 
and Ptolemy down to the end of antiquity. When the Finns re-appear 
in Prot. £, they seem to originate from Prot. Sk, through the medium 
of Prot. F, cf. § 24. The fact that the Finns are the sole representatives 
of the nations from Scandinavia in Prot. E, seems to show that Prot. Sk 
characterized them in a similar manner, as did Tacitus. 

i. Examination of Details. 

North: Finnoi = Finlanders. 

West: Chaideinoi = Heinir in HeiSmprk, now Hedemarken, Norway. 

East: *Souionai (Fauonai) = Swedes, in Upland. 
— : *Finaithoi (Firaisoi) = the Finaithae, in Finnhei5r, now Finveden. 
As they are placed in the east, we must assume that they occupied 
the entire space between the Baltic coast and the county of Fin- 
veden, that is to say: the present province of Smaland. The 
name survived on the western frontier owing to the ethnic con- 
trast to the Scandinavians. 

South: *Gautoi (Goutai) = Gotlanders. 

— : *Daneiones or *Dankiones (Daukiones) = Danes, in Scania; per- 
haps with suffix -k as in the Danish Fanniker, Manniker, Lolliker, 
Lyviker i. e. inhabitants of the islands Fan0, Man0, Laaland, 

Only the midland tribe, Leuonoi, cannot be identified with certainly; 
perhaps near Liongakoping (Linkoping). 

Some authors have connected them with the Liothida of Jordanis 
who are, however, in reality the inhahitants of the Scanian county of 

Cf. also the Leones, mentioned in the Old Engl, epical catalogue 
Widsith without definite localisation. 

The emendations Fauonai, Daukiones Z> *Souionai, *Daneiones (or 
*Dankiones) are necessary. It would have been impossible in a detailed 
list of tribes like Prot. Sk to omit mention of the Swedes, the only 
Scandinavian tribe of real Gothonic nationality noticed by Tacitus. And 
among some hundred Gothonic tribe-names, there is only a single one 
with the initial sounds Da-, viz. the Danes. Cf. our § 7. 

j. Conclusion. 

Prot. Sk may be called well verified both from topographic and 
linguistic points of view. 

138 Ptolemy's maps of northern Europe 

It is a most excellent piece of ethnic topography. The localisations 
are all correct. We notice especially the correct selection of names 
according to their statistical prominence. 


After finishing our survey of Ptolemaic prototypes, we reserve a 
separate paragraph for the question of limits which has been provision- 
ally mentioned in § 16, d. Cf. Figures 28, 29. 

Generally, it is taken for granted that Ptolemy represents the Cimbric 
Chersonese and the Scandian islands as Germanic without making any 
distinction from the area of the present Germany. He is again supposed 
to agree with his predecessors, Pliny and Tacitus, and the assumed com- 
mon scheme of these three authors is regarded as the classical norm. 

Only some few modern scholars interpret the classical evidences diffe- 
rently, introducing a scheme of distinction within the area of classical 
Germania. So e. g. Ad. van Kampen, in "Perthes' Atlas antiquus", 
1892, incorporates the Cimbric Chersonese with Germania, whereas the 
Danish islands and the Scandinavian peninsula are placed outside, design- 
ated as Germanic in a less pronounced degree. The map concerned re- 
appears unaltered in the 8th edition, 1908, published by Max Schneider. 
K. Wolff, in the 6th edition of Meyer's "Konversationslexikon", 1906, 
makes Germania embrace also the Danish islands, but still places the 
Scandinavian Peninsula apart. 

It must be admitted that those authors are mistaken who believe that 
Ptolemy represents Scandinavia as belonging to Germania without any 
restriction. The actual Ptolemaic distinction, however, differs radically 
from the schemes of the cartographers v. Kampen, Schneider, and Wolff. 

The northern frontier of the classical '^Germania proper", according 
to Ptolemy, does not exceed the limits of present Germany, -nay, of the 
Germanic Confederation before 1864. The Cimbric Chersonese and the 
Scandian islands are represented collectively, as a separate section. 

This appears from a series of various observations. 

I. Within the Ptolemaic text description of Germania, the Cimbric 
Chersonese is the only continental district which is represented separately. 
In other parts of Germania, the Ptol. constructor or his prototypes rather 
effaced existing sub-divisions. E. g., there is no trace of the Limes 
district, although it was occupied by the Romans, defended by strong 
frontier walls, and described in one of Ptolemy's special prototypes. — 
Instead of Bohemia, Ptolemy mentions a tribe of Bohemians, and corre- 


spondingly, the district of *Teurio-chaim has given rise to a so-called 
tribe Teurio-chaimai. 

2. The Cimbric Chersonese and the Scandian islands contain only- 
tribes, whereas the entire area of Germania proper contains in addition 
towns and other local details. This is a distinction, introduced arbitrarily 
by the Ptol. constructor. The Romans had visited the Cimbric Chersonese 
with their navy, and knew from practical observations details of this country, 
such as the headland Thastris (or Chartris), and the gulf Lagnus. On the 
other hand, the Roman armies and navies never visited what constitutes 
present Germany east of the middle and lower Elbe, and the Romans 
had no traceable connection whatever with the region between the Elbe 
and the Oder. Consequently, the Ptolemaic towns and rivers within the 
latter region must be regarded as fictitious. They are introduced by the 
Ptol. constructor, in order to produce the impression of homogenous 
geographical knowledge, embracing the entire area of "Germania proper". 
The Cimbric Chersonese and the Scandian islands are purposely repre- 
sented differently, as the less well-known periphery of the Ptolemaic 

3. Some of the oldest MS. atlases, viz. the Urbinas 82 and the Athous 
Vatopediensis, write the name "Kimbrike Chersonesos" with capital letters 
which are only a little smaller than those of the "Germania megale". 
(Noticed by J. Fischer). 

4. Some of the oldest MS. atlases, viz. the Athous Vatopediensis and 
the Burney 1 1 1 , represent the Cimbric Chersonese with colour, whereas 
the area of Germania proper is left blank. The Athos atlas extends the 
Cimbric colour also over the Scandian islands. It must be noticed that 
the two named MSS. represent both versions of the Ptolemaic atlas. 

We state: the only traceable boundary-line within classical Germania 
is the Ptolemaic which separates Germany from Denmark or Scandinavia. 


Our above investigations have given rise to a vast mass of hypo- 
theses within a field of study which has hitherto been scarcely cultivated. 

It is inevitable that such a first attempt will be productive of various 
errors, and we have already felt obliged to correct some mistakes, made 
in our previous sketches (in "The Scottish Geographical Magazine", etc.). 
We have also received letters from scholars who expressed doubts as to 
our results. 

It is now the part of the critics to reject our theory, or, if possible, 
to replace it by a better one. 


They may, e. g., try to reconstruct the assumable Ptolemaic proto- 
types on different lines, or to point out new and more decisive criteria. 

The investigation of corresponding prototypes within other parts of 
Ptolemy's atlas will also prove a practical means of verification. 

In face of all possible doubts and rejections, however, we venture to 
assume that one essential result has at any rate been obtained: the Pto- 
lemaic chaos is no more left completely without serious effort being made 
to dispell it. One attempt has now been made. 

Consequently, if geographers and ethnographers go on using the Pto- 
lemaic data frankly as ''positive" foundations, such as they have done for 
some five centuries, they will no more be able to excuse themselves with 
the absence of any genetic criticism. They will have to refute our statements, 
or to shrink anew from preserving and increasing the Ptolemaic chaos. 

Even this result will prove of considerable benefit. 

We hope that the eagerly expected publication of the Codex Urbinas 
82 through Jos. Fischer S. J. will attract the attention of scholars to 
this highly interesting, but also badly neglected branch of study, so that 
finally the chaos may be dissipated and the buried treasures of Ptolemy's 
predecessors become accessible and be duly utilised. 

Eskjcer pr. Jebjerg, Sailing, August i8^^' 1^14. 




Cf. Fig. 30 (designed after the printing of § 19). 

It deserves to be emphasized that Prof. L. Schmidt and the author 
of the present research have independently been led to the assumption of 
a prototype representing the physical map of Germany. We cannot indeed 
accept the traditional interpretation of the Ptolemaic mountains, as given 
by Schmidt: Melibokos = Harz, Semanus = Thiiringer Wald, Sudeta = 
Erzgebirge, Gabreta = Bohmer Wald, but the principal basis of agree- 
ment is at any rate worth comment. 

ad d. As we mentioned in our § 15, the Ptol. constructor seems to 
be fond of theoretical arrangements. One such is the Baltic coast-hne, 
running straight west-east, cf. § 20 c. We may add that the same theore- 
tical line west-east appears in the Melibokos, the Sudeta, and the Carpathian 
mountains; correspondingly, an inclination for a direction approximately 
north-south appears in the mountain Ketios south of the Danube, and in 
the rivers Vistula and Rhine. It need not be pointed out that such 
arrangements would chiefly affect the area of Prot. A, 

It is possible that both the Athos Map and the Burney Map reflect 
an original design in which the mountains were not so artificially modified 
as in the current Ptolemaic scheme. At least we notice that a pro- 
nounced oblique direction prevails in the Melibokos, according to both 
maps, and in the Sudeta, according to the Athos Map. 

Whereas Prot. A is probably not responsible for the horizontal and 
vertical lines of Ptolemaic mountains and rivers, we may, on the other 


hand, attribute to this prototype the exaggerated distance between the 
German frontier rivers and the mountains behind them. We notice the 
exaggeration east of the Rhine and north of the Danube. It is easily 
conceivable that the Romans were well informed concerning the regions 
directly contiguous with their frontier; and the large amount of known 
details from such regions would naturally tempt 20 cartographer to ex- 
aggerate the space concerned on the map. 

ad i. The so-called "town" Tulisurgion has in our § 20 c been com- 
pared with Tulifurdon in the vicinity, as a probable duplicate. Zeuss, 
"Die Deutschen", p. 7, suggests that Tulisurgion, Toulisurgion, is a mis- 
reading for *Teutiburgion, the famous wood in which the Romans under- 
went their fatal defeat in 9 A. D. — We now hold that his conjecture 
is correct, and therefore we have on our map Fig. 30 represented the 
vignette of the so-called "town" as a mountain which we attribute to the 
original prototype A. The vignette certainly occupies exactly the place 
of the mountain Teutoburger Wald, the present Osning. 

Orkynios, Lat. Hercynia, is a Celtic name meaning "wood" or "wooded 
mountain". Its primaeval Celtic form was *Percunia, corresponding to 
the Gothic word fairguni, "mountain". The original Hercynian Wood 
was a large complexe of middle German mountains and in mediaeval 
times the German form of the name -Fergunna, Vircunnia, etc. — still 
adhered to two distant chains, viz. i. Franken Hohe in Bavaria (probably 
a distortion: Franken for *Fergen); 2. Erzgebirge north-west of Bohemia. 
But the Ptolemaic Orkynios is neither of these; it must be the present 
Morayian Hills. At the southern extremity of this chain there is a 
mountain called Farren, which name seems to be a distortion of an 
ancient Gothonic *Ferhunja, a normal collateral form of *Fergunja 
according to Gothonic phonetic laws. 


§ 31. ADDITIONS TO § 22, PROTOTYPES Ac, Ad, Ae. 

On p. 82, we pointed out that the Tabula Peutingeriana has transplanted 
the words "Loci *VI regi(s)" from Dacia to Moesia, separating them from 
the continuation "Dae. Petoporiani". We suggested that the Tabula has 
correspondingly transplanted the town Sagadava = Zargidava Ptol., 
whereas Karsidava Ptol. would have been absorbed by the Moesian town 
Cahdava Tab. Further considerations have confirmed our suggestions, as 
we have discovered the Ptol. correspondence to one of the "loci regis 


Daci Petoporiani", viz. Piroboridava, read; *Piroporidava or *Pitoporidava. 
Ptolemy's b instead of / does not contradict the equation, as a similar 
shifting of media and tenuis appears in other Dacian names, cf. Biefoi = 
Piefigoi, Buridavensioi = Predavensioi, Potulatensioi = Polonda, Dierna = 

The Ptol. Piroboridava is placed in Moesia inferior, not far from the 
mouth of the Danube, but we must assume that the Ptol. constructor has 
displaced it too far towards the south-east together with Karsidava, 
whereas the accompanying town Zargidava was displaced towards the 
north-east. The Tabula Peutingeriana places the beginning of the words 
'^Dac. Petoporiani" north of the Carpathian mountains, and the Ptol. 
position of Karrodunon Ae = Karsidava Ad is corresponding. 

The Dacian king concerned, as we mentioned, appears in an inscrip- 
tion in Rome, Muratori 1039, 3^ "D- M- ^^^i Tiati fil. Dacae uxori Piepori 
regis Coisstobocensis Natoporus et Drilgisa aviae cariss. b. m. fecer." 
His people are obviously the Ptol. Koistobokoi in Roman Dacia, his 
residence is the Ptol. Piroboridava, and his other towns may be the 
neighbouring Tamasidava, Utidava, Trifulon, etc. We are informed by 
Dio Cassius LXXI, 12, that the Hasdings — a branch of the Vandals — 
invaded the country of the independent Koistobokoi about 172 A. D. He 
relates LXXII, 3, that a flock of 18,000 independent Dacians were about 
180 A. D. received in Roman Dacia. The place-names Piroboridava and 
Tamasidava through their forms betray a relatively late origin, as no 
other names on -dava are compounded with so long words. 

The combined evidence of Ptolemy, the Tabula, the inscription and 
Dio Cassius, affords a valuable piece of ethnic history dealing with north- 
eastern Dacia. We thus understand the relatively rich Ptolemaic descrip- 
tion of such peripheral parts of the Empire. 

The result is an interesting addition to our knowledge of ancient 
topography, but still more valuable is the statement that the edition of 
the Ptol. work can now definitely be dated as originating from after 
180 A. D., — a fact which we conjectured already from the occurrence 
of such tribal names as Biessoi and Sabokoi, cf. p. 89. 

In our genetic perspective p. 89, we ought perhaps to introduce a 
Ptolemaic stage VI, represented by the most freshly acquired informations 
such as the "loci VI regis Daci Petopori". The Post- Ptolemaic stage 
would then become nr. VII. 



The following lists do not pretend to offer an exact bibliography, — 
not even approximately. As no regular attempt at a Ptolemaic biblio- 
graphy has been made after 1837, it would be impossible to supply the 
want here, because the Ptolemaic statements are discussed and used in 
almost numberless works. It would be more than impossible during a 
time of European warfare, when visiting the libraries in the different 
capitals is prohibited. 

We therefore only try to point out some of ,the more important 
publications, hoping that it may prove useful to ordinary readers, and 
perhaps also contribute some practical hints to the work of a future 

A general bibliography down to the year 1837 is given by Heeren^ 
"Literatura Ptolemiaca". 


Famous, as it was, Ptolemy's Geography has been published in 
numerous editions or translations since the end of mediaeval times. We 
shall name some of the most important. 

1472. Latin translation, printed in Bologna. 

1533. Editio princeps in Greek by Erasmus, Basel. 

1838 — 1848. Edition with Latin translations by Wilberg, Essen. Con- 
tains the different readings of several MSS. 

1843— 1845. Ed. by Nobbe, Leipzig. Text-book for the practical use 
of scholars. 


1867. La Geographic de Ptolemee. Phototypical reproduction of the 
Mount Athos Manuscript, incl. the accompanying atlas, by Sewa- 
stionow and Langlois, Paris. The seven first pages of the atlas 
which were lacking, when the edition was published, have later 
been rediscovered in the British Museum by Jos. Fischer who is 
preparing an edition. 

1873. Edition of the chapters concerning Germany, Scandinavia and 
the neighbouring parts of Belgium and Sarmatia; in the "Germania 
antiqua" publ. by Mullenhoff, Berlin. Contains the different readings 
of several MSS. (arbitrary conjecture: Kyenones instead of 
Leuonoi !). 

1883 — 1 90 1. Edition with Latin translation and atlas by C. Muller, Paris 
(continued by Kurt Fischer \ Vol. Ill has not yet been published). 
Contains the different readings of most MSS. (except from the 
manuscript copies of the atlas). Valuable foot-notes. 


1484. Editio Ulmensis, with coloured atlas, designed by the famous 
German cartographer Nicolaus Bonis ^ Ulm. 

1490. Editio Romana, with atlas, Rome. Republished by Nordenskidld 
1889 (see below). 

1867. Atlas of the Mount Athos Manuscript, phototyp. reprod. by Se- 
wastionow and Langlois, Paris (see § 33). 

1889. A. Nordenskidld, "Facsimile- Atlas to the Early History of Car- 
tography" (Editio Romana), Stockholm. 

1 90 1. Tabulae XXX, a Ptolemaic atlas, reconstructed by C. Miiller, Paris 
(see § 33). 

1892. "Perthes' Atlas antiquus", Gotha, by Ad. v. Kampen. With re- 
constructed Ptolemaic map of the world. 8*^ edition, 1908. 

1900. Jelic, see § 35. With reproduced Ptol. map of Dalmatia from 
Cod. Urbinas 82. 

1 90 1. E. Devrient, "Hermunduren und Markomannen", in "Neue Jahr- 
biicher fur klassische Philologie". With reconstructed Ptol. map 
of Germania. 

1 90 1. R. V, Erckert, "Wanderungen und Siedelungen der germanischen 
Stamme in Mitteleuropa ... bis auf Karl den Grossen", Berlin. 
Monumental atlas, with reconstructed Ptol. map of Germania, based 
upon Miiller's edition. 

146 Ptolemy's maps of northern Europe 

1902. J. Fischer, **Entdeckungen der Normannen" = ^'The Discoveries 
of the Norsemen in America", London 1903. With reproduced 
Ptolemaic MS. maps. 

1904. A. Bjembo & Carl Petersen, ^'Claus Clauss0n Swart (Clavus)", in 
the "Danish Videnskabernes Selskabs Skrifter". German trans- 
lation 1909. With reproduced Ptol. MS. maps of Germania. 

1907. H, M. Chadwick, "The Origin of the English Nation". With 
reconstructed Ptol. map of Germania (p. 194 — 95). 

1 9 10. V. Novot7i<), in the publ. of the Bohemian Academy of Sciences 
(see § 35 a). With reconstructed map of Germania. 

191 1. R. Kiepert, "Formae orbis antiqui" (1894 — 1914 seq.). With re- 
constructed Ptol. map of Europe. 

191 1. Frithjof Nansen, "In Northern Mists". With reproduced Ptol. 

map of Europe and northern Asia from the Editio Romana. 
191 3 — 16. y. Fischer, see § 35 a. Reproduced Ptol. maps of the world, 

of Spain, Germania, and Scythia, from various MSS. 
191 5. A. Herrmann, Reconstructed map of Central Asia according to the 

scheme of Marinus. Cf. § 35 a. 


Under this heading, we try to point out some of the more important 
contributions to the general discussion of the theme, and besides some 
monographs dealing with the special topography of the Cimbric Cher- 

Valuable bibliopraphic collections concerning the Ptolemaic geography 
of Germania are contained in Novotnifs treatise "Ku kritice zprav Kl. 
Ptolemaia", 1910 (see below p. 147). 

a. Researches dealing with Ptolemy in a more or less general sense. 

1705 seq. J. A. Fabricius, "Bibliotheca Graeca". 3^^^ edition, Hamburg, 
1796; chapter dealing with Ptolemy V, 270 seq. 

1737. G. M. Raidel, "Commentatio critico-literaria de Claudii Ptolemaei 

1828. Heeren, "De fontibus geographicis Ptolemaei", in "Comment. 
Gotting.". VI, p. 66. 

1857. E, V. Wietersheim, "Ueber den praktischen Wert der speziellen 
Angaben in der Geographic des Claud. Ptolemaeus insbesondere 
liber Germanien", in "Berichte der sachsischen Gesellschaft der 
Wissenschaften", IX, p. 112 seq. 


1867. C. Muller^ "Rapports sur les manuscripts de la geographic de 

Ptolemee", in "Archives des missions scientifiques et litteraires", 

II'"^ serie, tome 4"^^. 
1881. — "Codex Vaticanus Nr. 191 der Geographie des Ptolemaeus", 

in "Hermes", XV. 
1 88 1. Th. Mommsen, "Zur Kritik der Geographie des Ptolemaeus", in 

"Hermes", XV. 
1888. Christ, "Geschichte der griechischen Litteratur", in Miiller's "Hand- 

buch der classischen Altertumskunde", 8*^ edition, 1905, VII, p. 506. 
1894. G. Holz, "Beitrage zur deutschen Altertumskunde; I. Ueber die 

germanische Volkertafel des Ptolemaeus", Halle. 
1894. Boll, "Studien iiber Claudius Ptolemaus", in Fleckeisen's "Jahr- 

biicher fiir classische Philologie". 
1897. R. Much, "Die Stadte in der Germania des Ptolemaus", in the 

"Zeitschrift fiir deutsches Altertum", XLI, p. 97 seq. 

Berger, "Die Grundlagen des Marinus-Ptolemaischen Erdbildes", 

in "Berichte der sachsichen Gesellschaft der Wissenschaften", hist. 

phil. CI., p. 87—143. 

A. Gnirs, "Das ostl. Germanien und seine Verkehrswege in der 

Darstellung des Ptolemaus", in "Prager Studien", IV. 

1900. y. Jelic^ "Das alteste kartographische Denkmal iiber die romische 
Provinz Dalmatien", in "Wissenschaftliche Mitteilungen aus Bosnien 
und Hercegovina", VII, p. 173 seq. 

1901. H. Zondervan, "Allgemeine Kartenkunde", Leipzig, 

1 90 1. W. Ketrzynski, "Die uns von Claudius Ptolemaus iiber Germania 
Magna iibermittelten Nachrichten", in "Publicationen der Akademie 
der Wissenschaften in Krakau, Anzeiger^ phil. Kl. 95." 

1 90 1. — "Kritische Bemerkungen iiber die Germania Magna des Ptole- 
maus", ibd. 8—15. 

1902. Ludwig Schmidt, "Zur Germania des Ptolemaeus", in Seeliger's 
"Historische Vierteljahrschrift", V, p. 79. 

19 10. V. Novotnii, "Ku kritice zprav Kl, Ptolemaia o zemich cesk^ch" 
in the Publications of the Bohemian Academy of Sciences. 

191 1. R. Kiepert^ commentaries upon the map XXIV etc. in the "Formae 
orbis antiqui". 

191 2. Jos. Fischer, S. J., "Die handschriftliche Ueberlieferung der Pto- 
lemaus-Karten", in "Verhandlungen des XVIII deutschen Geo- 

191 3. — • "An Important Ptolemy Manuscript with Maps in the New 
York Public Library", in the "Cathol. Hist. Records and Stu- 

1913- — "Die Strassburger Ptolemaus-Ausgabe vom Jahre 1513", in the 
"Stimmen aus Maria- Laach", Heft 3. 


148 Ptolemy's maps of northern Europe 

1913- Jos. Fischer, S. J., "Resultados de mis investigaciones cartograficas", 

in "Iberica, tirada del Segundo Congreso Espafiol' de' Geografia 

colonial y mercantil". 
191 4. — "El Valoso manuscrito latino de Ptolomeo de la universidade de 

Valencia". Ibd., Febr. 14**^. 
1914. — "Zur Ptolemausforschung" in "Petermanns Mitteilungen"; p. 287. 
1 9 16. — "Ptolemaus und Agathodamon". Mit einem Facsimile der Welt- 

karte des Agathodamon. Brit. Mus. Add. 19. 391". 

Publications of the Imperial Academy of Vienna, Vol. LIX, see 

under M2ik, -^Afrika". " 
1 91 3. P. Dinse, "Die handschriftlichen Ptolemauskarten", in the "Zen- 

tralblatt fiir Bibliothekswesen", XXX, p. 379 seq. 

191 3. — "Die handschriftlichen Ptolemauskarten und die Agathodamon- 
frage", in the "Zeitschrift der Gesellschaft fiir Erdkunde zu BerUn", 
p. 745 seq. 

19 1 4. K. Kretschmer, "Die Ptolemauskarten", in "Petermanns Mitteil- 
ungen", p. 142. 

, 1914. A. Herrmann, "Marinus, Ptolemaus und ihre Karten", in the 
"Zeitschrift der Gesellschaft fiir Erdkunde zu Berlin", 19 14, Nr. 10. 

191 5. — "Die Seidenstrassen vom alten China nach dem Romischen 
Reich", in the "Mitteilungen der k. k. Geographischen Gesellschaft 
in Wien", p. 472 seq. 

19 16. Hans V. Mzik, "Afrika der arabischen Bearbeitung der rscoyQacpixrj 
vcprjyYjOLg des Claudius Ptolemaus von Muhammad ibn Musa al- 
Hwarizimi. Herausgeg., iibersetzt und erklart von H. v. M. Kaiserl. 
Akad, d. Wissensch. in Wien, Philos. hist. Klasse, Denkschriften, 
LIX Bd., 4. Abh., Anhang II.". 

191 2. G. Schiitte, "Une carte du Danemark, de 1900 ans", in "Le Dane- 
mark", Oct. 

1913. — "A Map of Denmark, 1900 Years Old", in the "Saga-Book of 
the Viking Society", Vol.. VIII. 

' 1914 — 15. — "Ptolemy's Atlas, a Study of the Sources", in "The Scot- 
tish Geographical Magazine", with five continuations. 

1914. — "Der Ursprung der handschriftHchen Ptolemaus-Karten", in the 
"Mitteilungen zur Geschichte der Medizin und der Naturwissen- 
schaften", XIII, no. 5. 

191 5. — "Det ptolemaeiske Danmarkskort" , in "Geografisk Tidskrift", 
Hefte I. 

191 5, — "Danmarkskortet hos Ptolemaios ifolge Codex Burney 111", 
ibd. Hefte 2. 

191 5. — "Et maskeret Belgienskort hos Ptolemaios", ibd. Hefte 3. 

1916. — "Die Quellen der ptolemaischen Karten von Nordeuropa", in 
"Beitrage zur Geschichte der deutschen Sprache und Literatur", 
p. I seq. 


19 1 6. G. Schiitte, 'Ttolemaeiske Dubletter, Tripletter og Kvadrupletter", 

in '^Arkiv for nordisk filologi", XXXIII, p. 30. 
191 6. — "Nord og Mellemeuropa efter den rensede Ptolemaios", in 

"Geografisk Tidskrift", Hefte 7. 

b. Geographic or Ethnographic Compendia, etc. 

1822. N'. H. Brehmer, "Entdeckungen im Altertum", Heft i, p. 11. 
1837. C. Zeuss, "Die Deutschen und die Nachbarstamme". 2. edition, 

unaltered, 1903. 
1837. ^' Safarik (Schafarik), "Slavische Altertiimer" (translated from 

1852. W. defers, "Beitrage zur Geschichte und Geographie des alten 

Germaniens", Miinster u. Paderborn. 
1870 seq. K. Mullenhoff, "Deutsche Altertumskunde". Vol. II — V in 

posthumous edition. 

1877. J. Sadowski, "Die Handelsstrassen der Griechen und Romer, iibers. 
aus dem Polnischen von A. Kohn'\ 

1878. H. Kiepert, "Lehrbuch der alten Geographie", p. 10. 

1893. Berger, "Geschichte der wissenschaftlichen Erdkunde der Griechen", 

1899. O, Bremer, "Ethnographic der germanische Stamme", in Paul's 
"Grundriss der germanischen Philologie", 2. ed., Strassburg. Re- 
published in unaltered form 1905. Cf. our review in "Anzeiger 
fiir deutsches altertum", 1901. (Bremer on p. 825 points out a 
series of Ptolemaic duplicates). 

1899. Nystrom, "Geografiens och de geografiska upptackternas historia", 

1900. R. Much, "Deutsche Stammeskunde", in "Sammlung Goschen". 
2. ed. 1905. 

1904. K. Kretschmer, "Historische Geographie von Mittel-Europa'', in 
Below-Meinecke, "Handbuch der mittelalterl. u. neuer. Geschichte", 
Abt. IV. 
"190b. D. Detlefsen, "Ursprung, Einrichtung und Bedeutung der Erdkunde 
Agrippas", in Sieglin's "Quellen und Forschungen zur alten Ge- 
schichte u. Geographie", H. 13. 

1909. L. Schmidt, "Allgemeine Geschichte der germ. Volker bis zur 
Mitte des 6. Jahrh.", in Below-Meinecke, "Handbuch der mittel- 
alterl. u. neuer. Geschichte", Abt. II, 6. 

191 o. Jos. Fischer, S. J., "Die Entdeckungen der Normannen". 

191 1. Frithjof Nansen, "Nord i Takeheimen" = "In Northern Mists" 
(also in French and German). Christiania, London, etc. 

150 Ptolemy's maps of northern Europe 

c. Topography of the Cimbric Chersonese. 

1822. J. H.Bredsdorff, "Bidrag til Forklaring af Ptolemaei Efterretninger 

cm de nordiske Lande", in "Skandinaviske Litteraturselskabs 

Skrifter", XX. 
1836. E. C. Werlauff, "Bidrag til den nordiske Ravhandels Historic", in 

the Danish ''Videnskabernes Selskabs Skrifter", p. 271, 275. 
"1844. K. Miillenhoff, "Die deutschen Volker an Nord- und Ostsee in 

altester Zeit", in "Nordalbingische Studien", I. 
1868. P. Wislicenus, "Die Geschichte der Elbgermanen", Halle. 
1890. A. Erdmann, "Ueber den Namen und die Heimat der Angeln", 

in "Humanistiska Vetenskapsselskapets skrifter", Upsala. Reviewed 

by Herman M0ller, "Anzeiger fur deutsches altertum". Cf. our 

treatise "Var Anglerne Tyskere?", in "S0nderjydske Aarboger", 

1900, Flensborg. 
1894. y. F. Marcks, "Die romische Flottenexpedition zum Kimbernlande 

und die Heimat derKimbern", in "Jahrbuch des Vereins fiirAlter- 

tumsfreunde im Rheinland", Bonn. 
1899. Ihm, art. "Cimbri", in Pauly & Wissowa's "Realencyclopadie der 

classischen Altertumswissenschaften" . 
1904 — 1909. D. Detlefsen, "Die Entdeckung des germanischen Nordens 

im Altertum", in Sieglin's "Quellen und Forschungen zur alten 

Geschichte und Geographic". 
1907. H. M. Chadwick, "The Origin of the English Nation", in the 

"Cambridge Archaeological and Ethnological Series". Cf. our 

Review in "Arkiv for nordisk filologi", 1909. 
1912. R. W. Chambers, "Widsith", p. 241 etc., in the "Cambridge Ar- 

chaelogical and Ethnological Series". 

Further notice on existing literature may be found in the- works of 
Bremer, Novotny, Detlefsen, Nansen, and Chambers. 




Contents p. X (A) 

(cf. extract of contents - XXXIII) 






Fig. 2. 


Version A 

from the Codex Urbinas 82 in the Bibliotheca Vaticana, 13th century. 
By permission of Prof. Jos. Fischer S. J. Cf. § 19 — 20. 

-" ^-.'♦^l^ !! ' 'fc i 4? P ^ 



3 ~ .! < " 


"N M 

W^.."", ^K 

\ 2^p' 


1 ' ''^'^ W'^ 


Fig. 3. 


Version B 

from the Codex Bumcy 111, fol. 28, in the British Museum, 13th century. 

By permission of the Museum and "The Scottish Geographical Magazine". Cf. § ii — 20. 


Fig. 4. 

The type of the Roman editions 
from the Codex Ebnerianus (Lat.) in the New York Public Library- 
designed by Nicolaus Donis, 15th century. 

By permission of Prof. Jos. Fischer S. J. ("Catholic Historical Records and Studies' 
New York, 191 3, p. 222 — 223). Cf. § 19 — 20. 


Fig. 5. 

designed by L. Schmidt, in Seeliger's "Hist. Vierteljahrschrift", 1902, p. 84. 
By permission. Cf. § 19. 





i ^ 








IS o 


- 2 

c . 

(A O 

C U 

C/3 o 


o ^ 

" p 


a> ID 

a. > 



_ o 


Fig. 8. 


from the Codex Urbinas 82. 

By permission of Prof. Jos. Fischer S. J. and "The Scottish Geographical Magazine' 

Cf. § 21. 

Mountains, according to the Cod. Urbinas 82. 

Do. , according to the Cod. Burney, 1 1 : 







1^ Raitia 



Fig. 9. 

By permission of "The Scottish Geographical Magazine". Cf. § 21. 

Middle-German mountains, height 600 — 700 m. 
South-western mountains, height 1000 m. 

Space included by the Roman fortification lines and the district 
frontier of Raetia, corresponding to the Ptolemaic mountains 
Albia and Abnoba. 

Roman fortification lines. 

Frontier of the Roman province of Rsetia. 

Excavated Roman fortresses. 



am lyAris Flavis 


according to the Tabula Peutingeriana. 

By permission of "The Scottish Geographical Magazine". 
Cf. § 21. 

The Latin figures are road distances and indicate towns which 
belong to the Roman Limes district between the Rhine and 
the Danube. 


Fig. 11. 



By permission of "The Scottish Geographical Magazine". 
Cf. § 21. 




B i boKCTTv A I ki vnoems 

Ptolemy, Cod. Urbinas 82. 

CrinorioMC ScpUmiaci^u) 'BiricioniS ^ 

-i * i- ^ 

Tabula Peutingeriana. 

r*- iiDiriciArvis 

AUwo*,^ C 

Modem map. 

Fig. 12. DACIA 

from the Cod. Urbinas 82. 
By permission of Prof. Jos. Fischer S. J. Cf. § 27. 




vi: ,.:_- -.-L; 


<.ii- . 


(to left), contrasted with a modem physical map of the region of the Lower 

Danube (to right). 
By permission of "The Scottish Geographical Magazine". Cf. § 22. 

marks the Limit of Prototype Ae^ coinciding with the presumed southern Carpathian 
complexe of Ac. 






































(<J)T «i) 
























































Fig. 17. 


with the names redistributed according to their presumed correct positions. 




Names without brackets or in a single bracket ( ) are supposed to represent Ptolemy's pro- 
totypes Ac^ Ad, and Ae. Those within square brackets [ ] are names from the Tabula 
Peutingeriana; those within double brackets (( )) are names from Ptolemy's map of Ger- 
mania or from other sburces. 

Mercantile road^ from Carnuntum to Askaukalis, corresponding to Ptolemy's so-called 
"river Vistula". 


^ li 


H '-S 


s ^ 


U^ bx) 


^ o 


^ -5 


O 8 


Iz; c^ 


> ^ 

s ^ 

3 *^ 

CO g 


Fig. 19. 



Comparison of the duplicates. 
By permission of "The Scottish Geographical Magazine". Cf. § 23. 


































i XIX 

Fig. 20. 

contrasted with a modem map. 
By permission of "The Scottish Geographical Magazine". Cf. § 12. 

(in B2). Mountains, misinterpreted as a tribe or a town. 





:?: ■OT3 
O «= c 

Q O O 




O 3 


f^ ■§ 

C8 CO 

<i5 ^ 



O 3 

(S (S 








SuC^ rt 






. ^ c 



•§s § 




'0 > .H 
6^ 1 





"" c Pm 




s> « 


^ ^ 



0-5 - 
s-t II 






^^ = 











h «. — 










OJ ^ V 



J3 ^ 



H H 







Fig. 23. 

with the names from Prot. C. 

Cf. § 24. 





























, a 











• >M 





„ «« 


O ^ 




































^ c« 

;— I 


























































































>< -c 




rt C/3 











































Fig. 25. 


from the Cod. Urbinas 82. 

By permission of "The Scottish Geographical Magazine". Cf. § 26. 


('Rakotri'ai Si) 
('Hakatai BZ) 

CKotnoi BO 

Names in brackets, without addition of B or PI., belong to Prot. F. 
::•:;:;:: ^^6 area of antiquated names, borrowed from Herodotus. 



Fig. 26. 

with the names from the prototypes E and F. 
Cf. § 26. 

F I iN Nf'O J 







Names in brackets belong to Prot. E. 

Names without brackets belong to Prot. F, 

Names in square brackets are added from other sources (especially Pliny). 

..■..:...:....■: ^^^ ^""^^ ^^ antiquated names, borrowed from Herodotus. 


Fig. 27. 

Cf. § 27. 



Fig. 28. 

according to some modem representations, compared with the Ptolemaic map. 

Perthes' Atlas antiquus des. by A. van Kampen. K. Wolff, Germania. Meyer's Konversationslexikon. 

1892. 9th edition. 1916. 6th edition. 1907. 


according to some 13th century MSS. 

Cod. Urbinas 82. 
Version A. 

Cod. Athous Vatopediensis. 
Version A. 

Cod. Burney 111, 
Version B. 


Fig. 29. 


from the Codex Athous Vatopediensis, 13th century. 

By permission of The Danish Geographical Society. Cf. § 28. 

' *'"^'^^^^^,.^-'-*^' y--''^--'-''-\:f^f ^'•-^^■^'^'•'^''•'''•>4'i« ^A.-- 




Gothonic. Sound ch\ Charudes, Kauchoi, Chamauoi, Bainochaimai, 
Teuriochaimai, Chalusos. 

Termination -is\ Lirimiris, Marionis, Albis, Visurgis, (Amisis Mela 
= Amisias Ptol.), Marnamanis, Budoris, Kantioibis, Alkimoenis, Fur- 
gisatis, Kasurgis, Budorgis (duplicate of Budoris?), Limis, Askaukalis(?). 

Element -bergion, -burgion, "borough, barrow, mountain": Bergion, 
Askiburgion (town and mountain), Teutiburgion (town and ^mountain) ; 
-bokos, "beech": Melibokos; -chaim, "home": Bainochaimai, Teurio- 
chaimai; -maUy "men": Markomanoi; -vario, "men": Angrivarioi. 

Celtic. Termination -ak'. Mattiakon, Mokontiakon, Bibakon; -et: Sudeta, 
Gabreta, Nemetes. 

Element -briga, "borough": Artobriga; -dunon, "town": Lugo- 
dunon, Tarodunon, Segodunon, Eburodunon, Karrodunon, Noviodunon; 
■duron, "water": Batauoduron, Bragoduron, Boioduron; -magos, "plain": 
Borbetomagos, Noviomagos, Breukomagos; -lanion^ "place": Medio- 
lanion; -riton, "ford": Lokoriton; -carnon, "horn, rock": Karnus 

Pannonian. Sounds kv unaltered: Arsekvia, Akvinkon. (Gothonic alters 
kw into hwj Wy /, etc., whereas / appears in continental Celtic and 

Dacian. Element -^«z;«, "town": Setidava, Piroboridava, etc. 

Scythian. Termination -ss (o: s, sh): Pession, Trisson, Niosson; -an: 
Alanoi, Leianon, Kandanon, Bormanon. 


w Q 

C J) 



S Oh 



O ^ V 


o tuOrS, 


W o q « 


Plj « p^ 







Fig. 1. General Synopsis. 

- 2 — 4. Germania, Chersonesus Cimbrica, Scandia. 

5. Germania, Prototype J. 

6. Chersonesus Cimbrica and Scandia. 

7. North-western Germania, Chersonesus Cimbrica and Scandia, 
8 — II. South-western Germania. 

- 12 — 18. Dacia. 

- 19 — 20. The Mercantile Road from the Danube to the Mouth of the Vistula. 

- 21 — 23. Belgium and North-western Germania. 

- 24 — 26. North-eastern Germania and Sarmatia. 

- 27. Scandia. 

- 28 — 29. The Demarcation of Germania. 

• 30. A Rectified Ptolemaic Map of Nationalities. 

- 31. A Reconstructed Map of Nationalities. With lists of names denoting the various 

national types (p. XXIX). 


P. 145, § 34. First line; 1484, read: 1482. 

Third line; 1490, read; 1478, 1490. 



14 ^^1 Sh borrowed 

Renewed books aresubjecttoi"^ 


General Library ,